Electrical system with correction of errors in local voltage measurements and associated correction method

Abstract

The invention relates to an electrical system (3) comprising: an electrical bus (30) suitable for being electrically connected to electrical loads, electrical sources (31, 32, 33, 34) configured to supply the electrical bus (30) with respective powers, and local controllers (311, 321, 331, 341) respectively associated with the electrical sources, each local controller (311, 321, 331, 341) being configured to: receive a local voltage measurement acquired by a sensor associated with the local controller (311, 321, 331, 341); generate a local estimate (Viest) of a voltage (Vdc) of the electrical bus on the basis of the local measurement (Vimes) and power information; obtain a reference estimate (Vest_ref) of the voltage of the electrical bus, the reference estimate (Vest_ref) being common to all of the local controllers (311, 321, 331, 341), on the basis of the local measurement (Vimes), a reference setpoint (Vref) and a corrective term (dVi) representative of a difference between the local estimate (Viest) and the reference estimate (Vest_ref); generate a setpoint (Pci) for controlling the electrical source (31, 32, 33, 34) associated with the local controller (311, 321, 331, 341).

Description

[0001] DESCRIPTION

[0002] TITLE: Electrical system with local voltage measurement error correction and associated correction method

[0003] TECHNICAL FIELD

[0004] This disclosure relates to the general field of electrical systems, and in particular to the correction of voltage measurement errors in a decentralized control architecture.

[0005] STATE OF THE ART

[0006] Climate change is a major concern for many legislative and regulatory bodies worldwide. Indeed, various restrictions on carbon emissions have been, are being, or will be adopted by different countries. In particular, an ambitious standard applies to both new types of aircraft and those already in service, requiring the implementation of technological solutions to bring them into compliance with current regulations. Civil aviation has been actively working for several years now to contribute to the fight against climate change.

[0007] Technological research efforts have already led to very significant improvements in the environmental performance of aircraft. The Applicant takes into account the factors impacting all phases of design and development in order to obtain less energy-intensive and more environmentally friendly aeronautical components and products whose integration and use in civil aviation have moderate environmental impacts, with the aim of improving the energy efficiency of aircraft.

[0008] Consequently, the Applicant is constantly working to reduce its climate impact by employing sustainable development methods and minimizing greenhouse gas emissions. This sustained research and development work focuses in particular on new generations of aircraft engines and the development of electric propulsion technologies. An aircraft generally comprises an electrical network powered by electrical sources and is propelled by a turbomachine. To improve the aircraft's energy efficiency, it is possible to implement internal hybridization of the turbomachine. This involves adding a hybrid electrical system that forms an interface between the rotating mechanical shafts and the aircraft's electrical network.The hybrid electrical system aims to assist the turbomachine to optimize its operation, either by providing electrical assistance or by drawing power. It must manage the power distribution between the high-pressure and low-pressure mechanical shafts of the turbomachine, and in particular limit the power drawn, while ensuring that all electrical sources provide adequate power to the aircraft's electrical network. More specifically, in a multi-source architecture, generator-rectifier assemblies must supply the electrical network while controlling the power distribution between the turbomachine shafts, regardless of flight phases or load demands.

[0009] In a context of parallel power sources on a common power bus within a decentralized control architecture, local controllers regulate the voltage of the distribution busbar to share the power drawn from the different power sources. For example, a control architecture such as the one presented in W0202394751A1 is based on local droop control, allowing for adaptive power sharing constraints and prioritization of power sources to improve the stability and quality of the power grid.

[0010] Such solutions rely on measuring a network operating characteristic, typically voltage, locally—that is, for each local controller. However, they do not account for measurement errors. Thus, an error in the local measurement of a given source can lead to an error in the voltage regulation of that local source, and therefore a failure to maintain power sharing between the different electrical sources.

[0011] DESCRIPTION OF THE INVENTION

[0012] One purpose of this disclosure is to correct local voltage measurement errors to enable power sharing in a decentralized control architecture, within the general context of power sharing regulation in a multi-source electrical system.

[0013] This goal is achieved by an electrical system comprising:

[0014] • a clean electric bus designed to be electrically connected to electrical loads,

[0015] • electrical sources configured to supply the electrical bus with respective power levels, and

[0016] • local controllers respectively associated with the electrical sources, each local controller being configured to: - receive a local voltage measurement acquired by a sensor associated with the local controller,

[0017] - generate a local estimate of an electrical bus voltage from the local measurement and a power information;

[0018] - obtain a reference estimate of the electrical bus voltage, the reference estimate being common to all local controllers;

[0019] - from the local measurement, a reference setpoint and a correction term representing a difference between the local estimate and the reference estimate, generate a setpoint to control the electrical source associated with the local controller.

[0020] The proposed electrical system corrects the impact of local voltage measurement errors without modifying the existing decentralized control architecture. This architecture comprises local controllers, each responsible for controlling an associated power source based on the local voltage measurement. To achieve this, each local controller is configured to estimate the voltage of the power bus to be regulated, using the local voltage measurement and power information from its source. From these local estimates, a reference estimate is calculated. This reference estimate provides a common value calculated from the various estimates, and is therefore potentially more representative of the actual bus voltage. Each power source is then controlled by its local controller to supply power to the power bus, based on this common value.This ensures better power distribution between the different sources, as the setpoint takes this common value into account. Furthermore, the addition of a correction term allows for consideration of deviations in the local measurement from the reference estimate, with the control being calculated based on the reference setpoint, the local measurement, and the correction term.

[0021] The invention is advantageously complemented by the following features, taken individually or in any of their technically possible combinations:

[0022] - the setpoint is determined from a drastic curve depending on a drastic gain and an ordinate at the origin calculated from the correction term and the reference setpoint, the generated setpoint depending on the drastic curve so as to implement drastic regulation;

[0023] - each local controller is associated with a low-pass filter, each low-pass filter being configured to filter the local estimate into a filtered local estimate, and the reference estimate being obtained from the filtered local estimates;

[0024] - The reference estimate is a measure of central tendency of the local estimates respectively generated by the local controllers; each local controller is configured to receive the local estimate from the other local controllers via a communication bus and to obtain the reference estimate from the generated local estimate and the received local estimates, the communication bus having a sampling period between 1 ms and 100 ms. Each local controller includes a corrector configured to generate the correction term;The electrical system further includes a centralized controller separate from each local controller and connected to each local controller by an associated communication bus, the centralized controller being configured to receive the local estimate from each local controller via the associated communication bus, to generate the reference estimate and each correction term, and to transmit the associated correction term to each local controller; the correction term is a voltage differential between the reference estimate and the local estimate; which correction term is saturated by a maximum voltage correction; the correction term is stored by the local controller, so that in the event of a communication breakdown between the local controller and the local controllers of other electrical sources, the local controller is configured to generate the setpoint from the stored correction term;The electrical system further includes a network management system configured to transmit to each local controller information relating to the electrical sources connected to the electrical bus, each local controller being configured to obtain the reference estimate from the information relating to the connected electrical sources.

[0025] In another respect, a turbomachine is proposed that includes an electrical system as described previously.

[0026] According to another aspect, a method for correcting local voltage measurement errors on a dedicated electrical bus is proposed. This bus is designed to be electrically connected to electrical loads, with electrical sources configured to supply the bus with respective power levels. The error correction method comprises the implementation, by local controllers respectively associated with the electrical sources, of the following steps:

[0027] - reception of a local voltage measurement acquired by a sensor associated with the local controller, - generation of a local estimate of an electrical bus voltage from the local measurement and a power information;

[0028] - obtaining a reference estimate of the electrical bus voltage, the reference estimate being common to all local controllers;

[0029] - from the local measurement, a reference setpoint and a correction term representing a difference between the local estimate and the reference estimate, generation of a setpoint to control the electrical source associated with the local controller.

[0030] DESCRIPTION OF THE FIGURES

[0031] Other features, purposes and advantages of the invention will become apparent from the following description, which is purely illustrative and not limiting, and which should be read in conjunction with the accompanying drawings on which:

[0032] Figure 1 schematically illustrates an aircraft.

[0033] Figure 2 schematically represents a cross-sectional view of a turbomachine.

[0034] Figure 3 represents a hybrid turbomachine electrical system in an embodiment comprising four sources.

[0035] Figure 4 schematically illustrates the impact of a local voltage measurement error on the power distribution in a configuration comprising two sources.

[0036] Figure 5 is a flowchart of steps of an error correction process according to the invention.

[0037] Figure 6 represents a hybrid turbomachine electrical system in an embodiment comprising three sources and centralized error correction.

[0038] Figure 7 represents a local controller of an electrical source in the embodiment of Figure 6.

[0039] Figure 8 represents a hybrid turbomachine electrical system in an embodiment comprising three sources and distributed error correction.

[0040] Figure 9 represents a local controller of an electrical source in the embodiment of Figure 8.

[0041] Figure 10 schematically illustrates the impact of local voltage measurement error correction on the power setpoint generation in the two-source configuration. Throughout the figures, similar elements are labeled with the same reference numerals.

[0042] DETAILED DESCRIPTION OF THE INVENTION

[0043] In this application, upstream and downstream are defined with respect to the normal flow direction of the gas through the turbomachine. Furthermore, the longitudinal axis X of the turbomachine is the axis of rotation of its rotor parts. The axial direction corresponds to the direction of the longitudinal axis X, and a radial direction is a direction perpendicular to and passing through this axis.

[0044] By way of illustration, Figure 1 shows an aircraft comprising at least one propulsion unit 1, in this case two propulsion units 1. The present invention is not limited to an aircraft such as the one shown, but extends more generally to any other type of aircraft 100, such as a helicopter. The propulsion units 1 are attached and fixed to the aircraft 100, for example, each under a wing of the aircraft as illustrated, or alternatively on the wing of the aircraft or at the rear of the fuselage of the aircraft 100.

[0045] General overview of the electrical system

[0046] As is known, aircraft 100 comprises a plurality of electrical loads (or receivers) powered by electrical sources via an electrical network. More generally, an "electrical system" refers to a set comprising electrical sources and loads, connected to each other by the electrical network.

[0047] An electrical load is a device powered by electrical energy. This document focuses specifically on loads powered by a direct current network. It can be configured to transform the electrical energy that powers it into another form of energy, such as heat or mechanical energy.

[0048] Referring to Figure 3, the electrical loads can be divided into two categories. The first category, Ch1, is external to the propulsion system 1. These external electrical loads, Ch1, are specific to the aircraft 100 and may include, but are not limited to, an electric motor, a heating and / or air conditioning system, or a compressor. They enable a number of functionalities, whether the aircraft 100 is in flight or operating on the ground, such as pressurizing and / or illuminating the aircraft cabin, or operating the cockpit.

[0049] A second category of Ch2 loads is internal to the propulsion unit 1. The propulsion unit 1 generally includes a turbomachine 10, which will be described in detail later. The electrical loads Ch2 of the propulsion unit are loads of the turbomachine 10 (known as the "More Electric Engine" or MEE in English). For example, the electrical loads of the turbomachine Ch2 might include a starter, variable geometry components, or de-icing systems.

[0050] The plurality of electrical sources allows the electrical loads Ch1, Ch2 to be supplied with electrical energy.

[0051] On the one hand, with reference to Figure 2, the turbomachine 10 generally comprises two rotating bodies or shafts 181, 182, each associated with a respective generator motor forming an electrical source 31, 32, in order to supply the electrical network. In other words, the electrical source 31 is designed to draw power from a low-pressure body BP of the aircraft turbomachine 10, in order to supply direct current to an electrical bus 30 to which at least one electrical load Ch1, Ch2 is intended to be connected, and the electrical source 32 is designed to draw power from a high-pressure body HP of the aircraft turbomachine 10, in order to supply direct current to the electrical bus 30.

[0052] On the other hand, aircraft 100 includes a plurality of electrical sources 33, 34 distributed throughout aircraft 100, for example an Auxiliary Power Unit (APU), an electric battery, or a supercapacitor. These sources allow, in particular, for operation where the two rotating bodies 181, 182 of the turbomachine 10 are assisted, but also for supplying various onboard systems on the ground (pneumatic and hydraulic pressures, air conditioning, for example) when the propulsion units 1 are stopped in order to save fuel, or during start-up.

[0053] The electrical network typically comprises a set of electrical conductors, typically a set of cables linking together the different electrical loads Ch1, Ch2 and the different sources 31, 32, 33, 34, at the level of an electrical bus or distribution bar 30 of direct current.

[0054] Description of the turbomachine

[0055] Figure 2 schematically represents a cross-section of the propulsion unit 1 in a plane containing the longitudinal axis X. The propulsion unit 1 is intended to be mounted on the aircraft 100 via a pylon (not shown). The propulsion unit 1 comprises the turbomachine 10 and a nacelle 20 surrounding the turbomachine 10.

[0056] This disclosure is not limited to a twin-spool, twin-flow turbofan engine with direct fan drive 11 as illustrated. It extends more generally to various turbofan engine architectures, including unfaired turbofan engines, and other types of turbomachinery, which may have a different number of bodies and / or flows.

[0057] This disclosure falls more generally within the context of the internal hybridization of the turbomachine 10, i.e. presenting an electrical system 3 in interface with the mechanical shafts of the turbomachine and the electrical network of the aircraft 100, as detailed later.

[0058] The turbomachine 10 comprises, from upstream to downstream in the direction of the gas flow, a blower 11, a compression section 16 comprising a low pressure compressor 161 and a high pressure compressor 162, a combustion chamber 14, an expansion section 15 comprising a high pressure turbine 152 and a low pressure turbine 151, and an exhaust casing.

[0059] Each of the low-pressure compressor 161, high-pressure compressor 162, high-pressure turbine 152, and low-pressure turbine 151 comprises a rotor portion and a stator portion, the rotor portion being capable of being driven in rotation relative to the stator portion about the longitudinal axis X. The blower 11, the rotor portion of the low-pressure compressor 161, and the rotor portion of the low-pressure turbine 151 are connected to each other by a low-pressure shaft 181 extending along the longitudinal axis X, thus forming a low-pressure body (LP body), which is a first rotating body. The rotor portion of the high-pressure compressor 162 and the rotor portion of the high-pressure turbine 152 are connected to each other by a high-pressure shaft 182 also extending along the longitudinal axis X, around the low-pressure shaft 181, thus forming a high-pressure body (HP body), which is a second rotating body.As can be seen in Figure 2, the compression section 16, the combustion chamber 1, and the expansion section 15 are surrounded by an engine casing 21, to which are connected the stator parts of the low-pressure compressor 161, the high-pressure compressor 162, the high-pressure turbine 152, and the low-pressure turbine 151. The engine casing 21 defines a primary channel 13 allowing the flow of a primary gas stream A. A secondary channel 12 allows the flow of a secondary gas stream B.

[0060] As explained previously, the longitudinal axis X defines the axis of rotation for the blower 11, the rotor parts of the compression section 16 and the rotor parts of the expansion section 15, in other words for the LP body and the HP body.

[0061] The turbomachine 10 may also include at least one accessory gear box (AGB), typically housed in a cavity within the nacelle 20. The accessory gear box comprises a set of gears for rotating a plurality of shafts about their own axis, and accessories mounted on the plurality of shafts to derive useful mechanical power from their rotation. The gear set is itself driven by means of a radial drive shaft (RDS) connecting, possibly via a transfer case, the accessory gear box to at least one of the high-pressure (HP) and low-pressure (LP) units. For example, the radial drive shaft (not shown) may extend within a longitudinal cavity within an arm 17.In this way, mechanical power is likely to be taken from at least one of the high pressure body HP and the low pressure body BP to be delivered to at least one of the accessories via the accessory box.

[0062] Aircraft electrical system

[0063] Figure 3 illustrates an electrical system 3 distributed between the propulsion unit 1 and the aircraft 100 for supplying electrical power to external electrical loads Ch1 and / or internal electrical loads Ch2, typically by means of a direct current network.

[0064] As explained previously, the electrical system 3 makes it possible in particular to interface between the rotating bodies BP, HP of the turbomachine 10 and the electrical network of the aircraft 100. The electrical system 3 includes a first network specifically configured to meet the electrical power requirements of the external loads Ch1 and / or internal loads Ch2 of the propulsion unit 1 by mechanically drawing power from the turbomachine 10 via the electrical sources 31, 32, and a second network specifically configured to assist the start-up and / or in-flight operation of the turbomachine 10 using electrical sources 33, 34 of the aircraft 100.

[0065] In the embodiment illustrated in Figure 3, the electrical system 3 comprises four electrical sources. The low-pressure body BP is connected to an AC generator 31, coupled to the low-pressure shaft accessory housing 181. The high-pressure body HP is connected to an AC generator 32, coupled to the high-pressure shaft accessory housing 182. The electrical system 3 also includes a battery 33, which is a DC power source 33, and an auxiliary power unit 34.

[0066] More generally, this disclosure extends to any multi-source electrical system, including one or more alternating current electrical sources and / or one or more direct current electrical sources.

[0067] The alternating current generators 31, 32, 34 and the direct current generator 33 can each independently belong to the turbomachine 10, i.e., be controlled simultaneously with the turbomachine 10, or even be controlled by the turbomachine 10. Alternatively or complementaryly, the electrical sources 31, 32, 33, 34 can belong to the aircraft 100, i.e., be controlled simultaneously with the aircraft 100. In the embodiment of Figure 3, the electrical sources 31, 32 belong to the turbomachine 10, while the electrical sources 33, 34 belong to the aircraft 100.

[0068] The various electrical sources 31, 32, 33, and 34 are distributed throughout aircraft 100 at considerable distances, for example, several tens of meters. These distances can pose a problem for rapid communication between the electrical sources, as the capacity of the communication buses connecting the electrical sources 31, 32, 33, and 34 to the control system is distance-dependent, in terms of data rate and therefore the time required to transmit information. The communication buses can also be subject to electromagnetic interference from the electrical sources, electrical loads, or the various electrical devices in aircraft 100, particularly high-power equipment.Thus, it is advantageous to use an electrical system comprising several parallel sources on the electrical bus or distribution bar 30 with decentralized controllers or local controllers 311, 321, 331, 341, without communication between each electrical source 31, 32, 33, 34.

[0069] In the embodiment illustrated in Figure 3, the control system comprises a central "A / C control" unit, communicating with local control units. More specifically, the local HPS control unit communicates with the local controllers 311 and 321 associated with the BP and HP rotating bodies of the turbomachine 10, the BMS control unit communicates with the local controller 331 of the battery 33, and the APU CTRL control unit communicates with the local controller 341 of the auxiliary power unit 34.

[0070] Preferably, each electrical source 31, 32, 33, 34 is associated with a respective converter 310, 320, 330, 340, that is, with an element configured to supply electrical power. For example, the alternating current generator 31, 32, 34 is associated with an AC / DC converter 310, 320, 340 enabling the supply of direct current.

[0071] The electrical system 3 includes the electrical bus 30, connected to at least one of the external electrical load Ch1 and the internal electrical load Ch2 to the propulsion assembly 1, preferably to both types of loads Ch1, Ch2.

[0072] Furthermore, each of the converters 310, 320, and 330 is connected to the electrical bus 30. The electrical bus 30 may include a distribution linking the electrical sources 31 and 32 to the internal loads Ch2, and a distribution linking the electrical sources 33 and 34 of aircraft 100 to the external loads Ch1 specific to the aircraft. The electrical bus 30 is configured to supply electrical power or current to the electrical load Ch1 and Ch2 in the form of a continuous signal, in particular to meet the power requirements of the electrical load Ch1 and Ch2.

[0073] In fact, at least one, if not each, of the converters 310, 320, 330, 340 is configured to regulate the voltage of the electrical bus 30 from electrical power supplied by the respective power source 31, 32, 33, 34. The number and type of converters

[0074] 310, 320, 330, 340 and electrical sources 31, 32, 33, 34 is, of course, not exhaustive.

[0075] In order for the elements connected to the electrical bus 30, in particular the electrical loads Ch 1 , Ch2 to function correctly, it is necessary to regulate the voltage of the electrical bus 30.

[0076] To do this, each 310, 320, 330, 340 converter is connected to a local controller

[0077] 311.321.331.341 respectively. The decentralized control architecture can be a droop-type static regulation, for example as implemented in W0202394751.

[0078] Such a decentralized control architecture may be constrained by power sharing and withdrawal limits. However, this architecture relies on local voltage measurement, that is, measurement as close as possible to each power source. Using local measurement for each power source 31, 32, 33 makes it possible to detect the loss of a power source and to make the electrical system 3 more robust.

[0079] Preferably, each local controller 311, 321, 331, associated respectively with each electrical source 31, 32, 33, is linked to a sensor 51, 52, 53 configured to perform a local voltage measurement (Vimes), received by the local controller 311, 321, 331. The Vimes measurement corresponds to a local measurement, that is, one taken as close as possible to each electrical source. Performing a local measurement, as close as possible to each electrical source, is simpler in terms of installing the electrical system 3 in the aircraft 100 than having to route the same measurement to each local controller in the system.

[0080] In general, drastic control involves allowing the electrical bus 30 to have a bus voltage Vdc that can vary slightly between a maximum voltage Vmax and a minimum voltage Vmin, so that the electrical power supplied by each electrical source 31, 32, 33, 34 depends on the voltage Vdc. The maximum voltage Vmax corresponds to the bus voltage Vdc when there is no load.

[0081] The bus voltage Vdc is thus used to couple the power sources 31, 32, 33, 34 together, without requiring communication between them. More precisely, each power source 31, 32, 33, 34 is regulated to supply a current dependent on the AV bus voltage drop (equal to Vmax - Vdc). In this way, the regulation of the power sources 31, 32, 33, 34 can be carried out independently of each other, but they remain coupled by the bus voltage Vdc so as to reach a balance point together. For example, each respective local controller 311, 321, 331, 341 is designed to regulate the current supplied by its associated power source 31, 32, 33, 34 so that this supplied current follows a reference current calculated as a function of the AV voltage drop.

[0082] In other words, the "droop" type control law is based on a relationship, for each source, between the (actual) voltage Vdc of the electrical bus 30 and the current intensity to be supplied by the electrical source. As explained previously, an error in the local voltage measurement of a given source, for example due to an offset or a linearity error, leads to an error in the voltage regulation and therefore a failure to maintain power sharing between the different sources 31, 32, 33, 34.

[0083] This is illustrated in Figure 4, in the case of two distinct sources exhibiting two different control laws obtained from distancing curves. On the graph on the left, the two local measurements Vmes and V2mes are equal (and assumed to be correct and here equal to the actual bus voltage Vdc, assuming that voltage drops in the cables connecting the sources 31, 32 to the electrical bus 30 are neglected).

[0084] A power setpoint or controlled power Pc1, Pc2 is obtained for each source, as a function of the reference voltage Vref. On the graph on the right, the local measurement Vîmes is incorrect and exhibits an offset error ô mes with the local measurement V2mes always corresponding to the actual bus voltage Vdc in the simplifying assumption (Vîmes = Vdc+5 mes ). We observe that the error ô mesThis will cause the power distribution between the two sources to vary, resulting in a drift in the controlled power Pc1 for the electrical source 31, leading to a drift in the power distribution between the two sources: here, the first source 31 will supply more power than the second source 32, which is not the case in the fault-free operating mode illustrated on the left. The higher the droop coefficient, or droop gain, the lower the slope of the affine portion (or droop resistance, corresponding to the inverse of the droop gain) of the droop curve or control law; that is, the more horizontal the curves will be and the lower the offset error δ mes will result in a significant error in power sharing.

[0085] The reference setpoint Vref is assumed here to be equal for both sources 31, 32, and for example close to the nominal bus voltage V0. More generally, we can have distinct reference voltages Vrefl, Vref2 for each of the electrical sources 31, 32, so as to modify the power distribution between the sources 31, 32.

[0086] Hereafter, a quantity relating to an electrical source will be denoted by the subscript "i". Two configurations of the electrical system 3 in the case of three electrical sources 31, 32, 33 are illustrated in Figures 6 and 8. The subscript "i" here takes the values ​​of the integers 1 to 3.

[0087] The process of correcting voltage differences between the different electrical sources 31, 32, 33, 34 is schematically illustrated in Figure 5.

[0088] The process includes a step S1 of receiving the local Vimes voltage measurement acquired by the sensor 51, 52, 53 associated with the local controller 311, 321, 331.

[0089] In order to correct the local error, and more specifically to correct the voltage differences between the different electrical sources 31, 32, 33 impacting the respective control law, it is proposed to add to the voltage regulation at the level of each local controller 311, 321, 331 a correction term dVi specific to each electrical source 31, 32, 33.

[0090] During an S2 step, each local controller 311, 321, 331 generates a local estimate Vi estThe bus voltage Vdc of the electrical bus 30 is determined from the local Vimes measurement and a power information. Typically, the bus voltage Vdc corresponds to the actual voltage across the terminals of the electrical bus 30. Each electrical source 31, 32, 33 does not have access to the bus voltage value Vdc (for example, via a measurement), and must therefore estimate it.

[0091] Each local controller 311, 321, 331 can include an EST estimator implementing an estimator or a filter configured to generate the local estimate Vi est bus voltage, from the measured local voltage Vimes and power information.

[0092] The power information can be the power supplied by the electrical source 31, 32, 33. Preferably, the power information is the control value, i.e., the local setpoint or controlled power Pc1, Pc2. This avoids the need for additional sensors in the local controller 311, 321, 331. For example, with reference to Figures 7 and 9, the EST estimator of the local controller 311 takes as input the local bus voltage Vîmes measured by the sensor 51 associated with the source 31, and the local power Pc1.

[0093] Power information is not necessarily power, but can more generally be information relating to current, for example the current setpoint of the electrical source 31, 32, 33.

[0094] Power information can be the current measured locally, that is, near the electrical source. The power information used by the estimator EST to generate the local estimate Vi est This can be obtained by direct measurement. For example, the local controller 311, 321, 331, 341 may include a sensor, typically a current sensor. Current measurement is generally performed in power electronics. In this case, the power information may correspond to current measurements of the AC phases. As another example, the power information may be a DC current measurement of the electrical bus 30.

[0095] For example, in a configuration where the electrical system 3 includes a centralized control unit 40, configured to exchange information with the local controllers 311, 321, 331, the power information may correspond to a direct measurement of the bus voltage Vdc by a sensor 5. The local estimate Vi est can take into account, or even be equal to, the bus voltage measurement Vdc.

[0096] It will be understood that the bus voltage measurement Vdc is not a local measurement within the meaning of the invention. Hereafter, the term "estimated" (local or reference) always refers to the bus voltage being estimated. The bus voltage Vdc can be estimated using various techniques. For example, the local estimate Vi est can be generated by a Luenberger observer, configured to provide the local estimate Vi estfrom the measured local voltage Vimes and power information. In a manner known per se, the Luenberger observer allows the reconstruction of the state, here the bus voltage Vdc, from a model of the dynamic system and measurements of other quantities. The Luenberger observer includes a gain matrix determined such that the error in the state converges exponentially to 0.

[0097] For example, the local estimate Vi est can be generated by a Kalman filter, configured to provide the local estimate Vi est based on the measured local voltage (Vimes) and power information. In a manner known per se, the Kalman filter is an infinite impulse response filter that estimates the states of a dynamical system from a series of incomplete or noisy measurements. It comprises an evolution or transition matrix representing the dynamical system, and an observation matrix.

[0098] For example, the local estimate Vi est can be generated by a sliding mode observer, configured to provide the local estimate Vi est from the measured local voltage Vimes and the power information. The sliding mode observer consists of constraining, using discontinuous functions, the dynamics of the dynamic system to converge towards a manifold called the sliding surface, whose dimension is reduced by the dimension of the measurement vector. Preferably, the local estimate Vi est is generated by a model of the system; for example, the EST estimator implements a model of the system to calculate the local estimate Vi estBy "system model", we mean that the EST estimator can take into account a model of the power source 31, 32, 33, 34 to which the local controller is associated, a model of the cable leading to the reference point, i.e. the cable connecting the converter 310, 320, 330, 340 associated with the power bus 30. For example, the EST estimator can take into account a voltage drop in the cable connecting the power source associated with the power bus 30.

[0099] It will be understood that the estimated local Vi est is the voltage that the local power source expects to be present at the electrical bus 30. To determine the local estimate Vi est , we can thus rely on the knowledge of the measured local voltage Vimes and the voltage drop which we estimate from the knowledge of the current or the power.

[0100] Preferably, the system model also takes into account any filtering elements present on the electrical bus 30.

[0101] Preferably, for each local controller 311, 321, 331, the local voltage measurement Vimes can be filtered before being used by the EST estimator to calculate the local estimate. This advantageously reduces noise and spectral aliasing. Thus, the local estimate Vi est The Vdc voltage of the electric bus is generated from the filtered local Vimes measurement.

[0102] Preferably, for each local controller 311, 321, 331, the local estimate Vi est The Vdc voltage of the electrical bus is filtered. Typically, the local estimated Vi est can be filtered by a low-pass filter. In other words, the low-pass filter is configured to filter the local estimate Vi est in a filtered local estimate.

[0103] During step S3, a reference estimate Vest_ref of the voltage Vdc of the electrical bus 30 is obtained. The reference estimate Vest_ref is common to all local controllers 311, 321, 331, 341, and therefore to all electrical sources 31, 32, 33.

[0104] More specifically, the local estimates Vi estThe values ​​generated by each local controller 311, 321, and 331 are then transmitted to a voltage measurement offset controller 41. The controller 41 is configured to determine a reference estimate Vest_ref of the bus voltage Vdc, which is intended to be as representative as possible of the actual bus voltage Vdc. The reference estimate Vest_ref is the best estimated bus voltage. It can be determined by averaging the local estimates or can be directly measured by a voltage sensor on the bus. As will be detailed later, the reference estimate Vest_ref can be obtained by a centralized controller 41 communicating with each local controller 311, 321, and 331, or by decentralized controllers 41 included within each local controller 311, 321, and 331.

[0105] Based on local estimates 7i estGiven the bus voltage associated with the different sources 31, 32, 33, the reference estimate Vest_ref can be obtained in several ways. Generally, the reference estimate Vest_ref is calculated from the different local estimates, for example as a statistical quantity derived from the different local estimates, preferably a measure of central tendency of the local estimates Vi est respectively generated by local controllers 311, 321, 331, 341.

[0106] For example, the reference estimate Vest_ref could correspond to the average of the local estimates. In other words, considering that we have electrical sources, we would have

[0107] Vest

[0108] Alternatively, the reference estimate Vest_ref can correspond to the median of the local estimates Vi est , that is, Vest_ref = median(71 est , V2 est , ..., Vn estAlternatively, the reference estimate Vest_ref can correspond to a weighted average of the local estimates Vi est , the weightings may depend on the reliability of the local Vimes measurement of each local controller 311, 321, 331, 341 for example.

[0109] The reference estimate Vest_ref can be calculated from the filtered local estimates.

[0110] Preferably, the reference estimate Vest_ref calculated from the local estimates Vi estThis can be filtered, typically by a low-pass filter. Indeed, the error in the local measurement that we seek to correct evolves slowly compared to the overall dynamics of the electrical system. It is therefore relevant to remove high-frequency variations from the reference estimate Vest_ref in order to retain only components representative of the error between the different electrical sources, for example, a drift of the sensor 51, 52, 53 performing the local measurement. This prevents the correction from impacting the stability of the voltage regulation. Preferably, the filter used has a lower bandwidth than the bandwidth of the voltage regulation. In other words, the correction of the reference estimate Vest_ref is performed with a slower dynamic range than the bandwidth of the voltage regulation, preferably 10 times slower, for example, between 10 Hz and 500 Hz.

[0111] Preferably, filtering is performed within each local controller 311, 321, 331. This mitigates the effect of variation due to sampling or quantization. During an S4 step, each local controller 311, 321, 331, 341 generates a setpoint Pci to control the associated power source 31, 32, 33, 34. The setpoint is generated from the local measurement Vimes, the reference setpoint Vref, and a correction term dVi representing a deviation from the local estimate Vi est and the reference estimate Vest_ref.

[0112] More specifically, once the reference estimate Vest_ref is calculated, the corrector 41 estimates the correction term dVi specific to each source 31, 32, 33, 34 from the local estimate Vi estand the reference estimate Vest_ref. The dVi correction terms correct voltage errors in the electrical sources, indirectly preventing power sharing drift. More specifically, this allows any overall voltage error to be distributed among the different sources, thus ensuring a balanced distribution of power supplied by the various electrical sources to the electrical bus 30.

[0113] Typically, the correction term dVi can be the voltage deviation, that is, the difference, between the reference estimate Vest_ref of the bus voltage Vdc and the local estimate Vi est In this case, the local voltage measurement can be corrected during the generation of the Pci setpoint by setting dVi = Vest_ref - Vi est .

[0114] Preferably, the correction term dVi is saturated at the maximum voltage correction S opossible in both positive and negative values. More precisely, corrector 41 includes a saturator such that dVi = sign(Vest_ref — Vi est x min(50, |Vest_ref — 7i est |). Typically, the maximum voltage correction S o is low compared to the nominal voltage V0 of the electrical bus 30 or the voltage setpoint Vref, around which the bus voltage Vdc is to be regulated. For example, the maximum voltage correction S o The value is less than 10% of the nominal voltage V0, or even less than 5% of the nominal voltage V0. Saturation prevents drift in situations where a local estimate would begin to drift, for example, due to an error in the local Vimes measurement caused by a failure of the sensor 51, 52, 53 associated with the power source. Saturation can be defined as the maximum error of the sensor 51, 52, 53, typically between -2% and +2%.

[0115] The correction term dVi can then be passed to each local controller (Figure 7) or calculated within each local controller 311, 321, 331, 341 (Figure 9). As illustrated, each local controller can include its own REG, so that the corrector 41 outputs the correction term dVi to the REG of local controller 311. The correction term dVi can be calculated from the filtered reference estimate and / or the filtered local estimates. Preferably, the correction term dVi can be filtered, typically by a low-pass filter.

[0116] Preferably, each local controller 311, 321, 331 stores a reference setpoint value Vref in memory. The stored value can be identical for all power sources 31, 32, 33. Alternatively, the reference setpoint Vref can be updated, i.e., modified dynamically, for example by a central control system of the power system 3. In the case of "drop" type control, the reference setpoint Vref can correspond to the nominal bus voltage VO of the power bus 30.

[0117] Each local controller 311, 321, 331 can then generate the Pci setpoint to control the associated power source 31, 32, 33, 34, from the local measurement Vimes, the reference value Vref and the correction term dVi.

[0118] For example, in the case of "droop" control, the setpoint generation depends on a droop gain and an y-intercept that are themselves dependent on the correction term dVi and the reference setpoint Vref. The generated setpoint, or controlled power, depends on the droop curve and the actual bus voltage Vdc, thus droop regulation is implemented. Taking a local measurement corrected by the correction term in the droop regulation improves power sharing between the different electrical sources and prevents the electrical source from deviating from the expected behavior.

[0119] The effects of the proposed correction on the droop control are illustrated in Figure 10. Here, the y-intercept of the curves relating the actual bus voltage Vdc and the controlled power Pc1, Pc2 to each source 31, 32 is modified by adding the correction term dV1, dV2. This results in a new power distribution between the two sources P'c1, P'c2 that is closer to the error-free distribution. The second source 32 again provides more power than the first source 31, as in the error-free situation shown in Figure 4.

[0120] The REG controller can determine the Pci setpoint power which is transmitted to an internal control loop CONT of the associated source, the internal control loop CONT being intended to transmit a power drawdown setpoint to the converter of the associated source.

[0121] For example, with reference to Figures 7 and 9, the REG controller takes as input the error e v estimated that the regulator REG is seeking to correct. The error e v corresponds to the sum of the difference between the reference setpoint Vref and the local voltage measurement Vîmes, and the correction term dVi. Thus, we have v = Vref - Vîmes + dVi.

[0122] In this embodiment, the drastic gain specific to each source is constant. In another embodiment, the drastic gain for the regulation of each electrical source 31, 32, 33 can be dynamic, that is, modified by the central regulation system of the electrical system 3.

[0123] Thus, if one of the sensors provides a local voltage measurement Vîmes with an error, for example, an error of 2% of the nominal bus voltage Vdc value VO, the power sharing between the different power sources 31, 32, 33 is impacted by the measurement error, in the absence of the addition of the correction term dVi, or in other words, without voltage compensation. For example, if the voltage measured with error is higher than the actual bus voltage and the reference voltage Vref as illustrated, the power generated by the associated source will be higher than that of the other sources. This leads to an increase in the actual bus voltage. Adding the correction term dVi at the level of each local controller 311, 321, 331 addresses the power sharing problem and leads to a better power distribution, typically an equal distribution, between the three power sources 31, 32, 33.

[0124] In order to be able to exchange local estimates Vi est bus voltage, communication is implemented between each electrical source 31, 32, 33, 34. Information can be exchanged via a communication bus in a distributed manner or can be sent to a centralized control unit 40.

[0125] Communication can be slow, typically with transmission times ranging from milliseconds to hundreds of milliseconds. This is because the dynamics driving local errors are very slow compared to the system's overall dynamics. In other words, the local errors we want to correct (offsets, errors due to poor linearity) vary slowly over time. Similarly, the estimation dynamics during step S1 can therefore be performed at a low frequency.

[0126] Example of a centralized architecture

[0127] In an embodiment illustrated in Figures 6 and 7, information exchange is centralized. In this embodiment, the electrical system 3 includes a centralized controller 41, separate from each local controller (311, 321, 331, 341), and configured to generate the reference estimate Vest_ref, and / or each correction term dVi. For example, the electrical system 3 includes a control unit 40, for example a computer, implementing the controller 41. The control unit 40 communicates with the various electrical sources 31, 32, 33, 34, and more specifically transmits and receives information from each local controller 311, 321, 331, 341. This embodiment can be advantageous if the network architecture already includes the control unit 40 or the computer, for example within a centralized control system configured to transmit power instructions to the local controllers 311, 321, 331.

[0128] In this embodiment, each local controller 311, 321, 331, 341 can be connected to the centralized controller 41 via an associated communication bus. The communication buses do not require a high sampling rate, i.e., high-frequency information exchange. Typically, communication buses have a sampling period between 1 ms and 100 ms. Thus, the centralized controller 41 is configured to receive the local estimate Vi estFrom each local controller 311, 321, 331, 341, via the associated communication bus, the reference estimate Vest_ref is generated from the local estimates, and the associated correction term dVi is derived for each power source 31, 32, 33, 34. The centralized controller 41 can then transmit the associated correction term dVi to each local controller 311, 321, 331, 341, allowing each local controller 311, 321, 331, 341 to calculate the corrected setpoint and control the associated power source, ensuring adequate power distribution.

[0129] Preferably, in this embodiment, the electrical bus 30 can be associated with a sensor 5 that measures a bus voltage across the terminals of the electrical bus 30 at a reference point. The reference estimate Vest_ref can then correspond directly to the voltage measured by the sensor 5.

[0130] In the event of a loss of communication with the centralized control unit 40, the corrector 41 no longer transmits information about the other electrical sources to each power source 31, 32, 33. Preferably, the last value of the correction term dVi is stored within the local controller 311, 321, 331 and used for generating the PCI command until communication with the centralized control unit 40 is restored.

[0131] Example of a distributed architecture

[0132] In an alternative embodiment, illustrated in Figures 8 and 9, information exchange is distributed. In this embodiment, a communication bus 42 is available between the electrical sources 31, 32, 33, and 34. More precisely, the communication bus 42 of the electrical system 3 connects the various local controllers 311, 321, 331, and 341 specific to each electrical source. Thus, each local controller 311, 321, 331, and 341 receives the local estimates Vi est calculated by the other local controllers. In other words, each local controller 311, 321, 331, 341 is configured to receive the local estimate Vi est from other local controllers via communication bus 42 and to obtain the reference estimate Vest_ref from the local estimate Vi estgenerated and local estimates received. As before, the communication bus 42 can have a low sampling frequency, for example a sampling period of between 1 ms and 100 ms.

[0133] In the case of a distributed architecture, if communication with a power source 31, 32, 33 is lost, the reference estimate Vest_ref can still be defined by removing the power source whose communication is lost until communication is restored. Typically, the reference value is then calculated from the n-1 available measured local voltages.

[0134] The local controller associated with the failed source no longer receives the local voltage estimates associated with the other sources. Preferably, the dVi correction term is stored by local controllers 311, 321, 331, and 341, so that in the event of a communication failure between one local controller and the other local controllers, the disconnected local controller is configured to generate the PCI setpoint from the stored dVi correction term. For example, the last calculated dVi correction term value is retained by local controllers 311, 321, and 331 and transmitted to the REG controller until communication with the local controllers of the other sources is restored.

[0135] In general, the configuration of the electrical system 3 can lead to situations where certain electrical sources are no longer connected in parallel on the same communication bus 42. Typically, when a contactor is opened or the electrical system 3 is reconfigured, communication with one of the electrical sources 31, 32, or 33 may be interrupted. Preferably, in this case, the disconnected sources are not taken into account by the controller 41 in the calculation of the reference estimate Vest_ref.

[0136] For example, each local controller 311, 321, 331 can store information about the electrical network configuration. Preferably, the electrical system 3 includes an Electrical Network Management System (ENMS) configured to transmit information about the electrical sources 31, 32, 33 connected to the electrical bus 30 to the local controller 311, 321, 331. The network management system is typically configured, for example, to control the contactors and thus knows the configuration status of the electrical system 3, in particular, which electrical sources are connected to each power source. The local controller 311, 321, 331 can then be configured to obtain the reference estimate Vest_ref from the information about the connected electrical sources 31, 32, 33, 34. Typically, this is done by decreasing the number of sources n in the calculation of the reference estimate Vest_ref as an average.

[0137] The network management system can be specific to each local controller 311, 321, 331 or be part of the electrical system 3 and communicate with the local controllers 311, 321, 331 in a centralized manner.

[0138] In the illustrated embodiments, the method for determining the correction term is applied to the electrical system 3 comprising three or four electrical sources 31, 32, 33, 34. More generally, the method can be generalized to any known but arbitrary number n of electrical sources connected in parallel on a DC electrical bus of the same voltage level. The type of electrical sources used in the architecture or in the electrical system 3 is also generalizable. The described method is not limited to reversible sources and can be extended to non-reversible sources. In particular, the number of electrical machines 31, 32 mounted on the propulsion unit 1 is not limited to two. Similarly, the electrical storage source is not limited to the battery 33 alone but extends to other types, notably a supercapacitor, and to a greater number.

[0139] The described method can also be applied to all types of decentralized control. In particular, it can be applied to a symmetrical decentralized control architecture, for example the power generation system presented in application FR3141678À1.

[0140] Finally, the described process is not limited to the context of the aircraft 100 electrical system 3, but can be applied to any electrical system comprising a plurality of electrical sources connected in parallel with decentralized control, for example a terrestrial electrical network, typically an intelligent network (or "smart grid").

Claims

DEMANDS 1. Electrical system (3) comprising: • an electric bus (30) suitable for being electrically connected to electric loads, • electrical sources (31, 32, 33, 34) configured to supply the electrical bus (30) with respective powers, and • local controllers (311, 321, 331, 341) respectively associated with the power sources, each local controller (311, 321, 331, 341) being configured to: • receive a local voltage measurement acquired by a sensor associated with the local controller (311, 321, 331, 341), • generate a local estimate (Vi est ) of a voltage (Vdc) of the electrical bus from the local measurement (Vimes) and a power information; • obtain a reference estimate (Vest_ref) of the electrical bus voltage, the reference estimate (Vest_ref) being common to all local controllers (311, 321, 331, 341); • from the local measurement (Vimes), a reference setpoint (Vref) and a correction term (dVi) representing a difference between the local estimate (Vi est ) and the reference estimate (Vest_ref), generate a setpoint (Pci) to control the electrical source (31, 32, 33, 34) associated with the local controller (311, 321, 331, 341).

2. Electrical system according to claim 1, wherein the setpoint is determined from a drastic curve depending on a drastic gain and an ordinate at the origin calculated from the correction term (dVi) and the reference setpoint (Vref), the generated setpoint depending on the drastic curve so as to implement drastic regulation.

3. An electrical system according to any one of claims 1 and 2, wherein each local controller (311, 321, 331, 341) is associated with a low-pass filter, each low-pass filter being configured to filter the local estimate (Vi est) in a filtered local estimate, and the reference estimate (Vest_ref) being obtained from the filtered local estimates.

4. An electrical system according to any one of claims 1 to 3, wherein the reference estimate (Vest_ref) is a measure of central tendency of the local estimates (Vi est respectively generated by the local controllers (311, 321, 331, 341).

5. An electrical system according to any one of claims 1 to 4, wherein each local controller (311, 321, 331, 341) is configured to receive the local estimate (Vi est from other local controllers via a communication bus and to obtain the reference estimate (Vest_ref) from the local estimate (Vi est generated and local estimates received, the communication bus having a sampling period between 1 ms and 100 ms.

6. Electrical system according to any one of claims 1 to 5, wherein each local controller (311, 321, 331, 341) includes a corrector (41) configured to generate the corrector term (dVi).

7. An electrical system according to any one of claims 1 to 5, comprising a centralized controller (41) separate from each local controller (311, 321, 331, 341) and connected to each local controller (311, 321, 331, 341) by an associated communication bus, the centralized controller (41) being configured to receive the local estimate (Vi est from each local controller (311, 321, 331, 341) via the associated communication bus, to generate the reference estimate (Vest_ref) and each correction term (dVi), and to transmit to each local controller (311, 321, 331, 341) the associated correction term (dVi).

8. An electrical system according to any one of claims 1 to 7, wherein the correction term (dVi) is a voltage differential between the reference estimate (Vest_ref) and the local estimate (Vi est ).

9. Electrical system according to any one of claims 1 to 8 wherein the correction term (dVi) is saturated by a maximum voltage correction (50).

10. Electrical system according to any one of claims 1 to 9, wherein the correction term (dVi) is stored by the local controller (311, 321, 331, 341), such that in the event of a communication breakdown of the local controller (311) with the local controllers (321, 331, 341) of the other electrical sources (32, 33, 34), the local controller (311) is configured to generate the setpoint (Pci) from the stored correction term (dVi).

11. Electrical system according to any one of claims 1 to 10, further comprising a network management system configured to transmit to each local controller (311, 321, 331) information relating to the electrical sources (31, 32, 33, 34) connected to the electrical bus (30), each local controller (311, 321, 331) being configured to obtain the reference estimate (Vest_ref) from the information relating to the electrical sources (31, 32, 33, 34) connected.

12. Turbomachine comprising an electrical system according to any one of claims 1 to 11.

13. Method for correcting local voltage measurement errors of an electrical bus (30) suitable for being electrically connected to electrical loads, electrical sources (31, 32, 33, 34) being configured to supply the electrical bus (30) with respective powers, the error correction method includes the implementation, by local controllers (311, 321, 331, 341) respectively associated with the electrical sources (31, 32, 33, 34), stages of - reception of a local voltage measurement (Vimes) acquired by a sensor associated with the local controller, - generation of a local estimate (Vi est ) of a voltage (Vdc) of the electrical bus (30) from the local measurement (Vimes) and a power information; - obtaining a reference estimate (Vest_ref) of the voltage (Vdc) of the electrical bus (30), the reference estimate (Vest_ref) being common to all local controllers (311, 321, 331, 341); - based on the local measurement (Vimes), a reference setpoint (Vref) and a correction term (dVi) representing a difference between the local estimate (Vi est) and the reference estimate (Vest_ref), generation of a setpoint (Pci) to control the electrical source (31, 32, 33, 34) associated with the local controller (311, 321, 331, 341).

Images