METHOD AND SYSTEM FOR CONTROLLING AN ELECTRIC MACHINE DRIVED BY AN INVERTER WITH MULTIPLE SWITCHING ARMS

Description

technical field

[0001] The present invention relates to a method for controlling a rotating electrical machine, preferably a synchronous electrical machine, for example of the synchro-reluctant or permanent magnet synchronous type.

[0002] The invention also relates to a control system implementing such a method.

[0003] It is known to employ rotating electrical machines such as permanent magnet synchronous machines or synchronous reluctance machines, particularly permanent magnet-assisted synchronous reluctance machines. Such electrical machines are used, for example, in the field of propulsion, for instance, to generate engine torque on board a vehicle such as a motor vehicle.

[0004] Document FR 3 051 296 A1 describes, for example, a synchro-reluctant machine assisted by permanent magnets.

[0005] A method for controlling such an electrical machine generally includes the calculation of so-called "direct" and "quadratic" (also called "quadratic") currents and voltages, which are currents and voltages expressed in a rotating frame linked to the rotor, and the implementation of two control systems, one dealing with the so-called "direct" quantities, and the other dealing with the so-called "quadratic" quantities, in order to determine the voltages to be applied to each phase of the rotating machine.

[0006] Therefore, determining the reference, direct, and quadrature components of the stator current is a crucial step at this stage, in order to guarantee optimal performance of the electrical machine. The complexity of determining the reference current depends on the nature of the machine used. For a smooth-pole synchronous machine, the direct and quadratic components of the current can be obtained in a linear and direct manner from the required torque, and for other machines, such as synchronous-reluctant machines, obtaining the reference components of the stator current requires more sophisticated and complex methods. Previous technique

[0007] The choice of a method for determining direct and quadrature current setpoints depends on the desired performance level and the real-time computational volume; a topic that has been extensively covered in industrial and academic literature. In this context, the following methods can be mentioned: a) The Maximum Torque Per Ampere (MTPA) method determines the direct (Id) and quadrature (Iq) components of the current using an optimal solution for the stator current vector angle. It is used when the voltage amplitude is below the maximum threshold. A flux reduction phase (flux reduction refers to the reduction of flux to increase the rotational speed of the electrical machine beyond its nominal value without increasing the supply voltage), achieved by applying the error feedback to the amplitude of the reference voltage, is integrated when the voltage exceeds the maximum threshold. In this respect, the literature distinguishes two ways of implementing flux reduction through the stator current angle: voltage error feedback, which acts directly on the direct current component (Id), and voltage error feedback, which acts on the current vector angle.b) A method in which the current components Id and Iq are determined from offline data tables obtained from optimal torque and flux methods and models. A defluxing phase via voltage error feedback can be integrated when its voltage amplitude exceeds a maximum threshold. c) A method in which the current components Id and Iq are determined by MTPA and a single-component current regulator (often the forward component Id) to define the voltage angle in the defluxing region. d) A method in which the current components Id and Iq are determined by simplified real-time optimization algorithms of a model based on inductances independent of the stator current components.

[0008] Other examples of methods and systems for controlling synchronous electric machines are described in particular in patent applications US 9479102, US 2017 / 085200 A1, US 9768719, US 7586286, US 6936991 and US 9614473.

[0009] US patent application 9479102 describes a system and method for controlling a synchronous electric machine. The control of the synchronous electric machine is based on a torque setpoint and an estimate of the magnetic flux associated with the rotor. This control allows for flux defluxing of the electric machine. However, such a method may not be robust to the DC bus voltage limit.

[0010] Furthermore, US patent applications 2017 / 085200 A1 and 9768719 describe methods and systems for controlling synchronous electric machines with permanent magnets. The control of the electric machine is based on MTPA (Maximum Torque Per Ampere) tables. Moreover, the control modifies the current setpoints according to the voltage limit of the electric machine. Consequently, by modifying the current setpoints after using the MTPA tables, this control does not guarantee the accuracy of the voltage limit determination, which can lead to over- or under-utilization of the inverter bus voltage.

[0011] In US patent applications 7586286 and 6936991, the determination of current components Id and Iq is based on a Peak Torque Per Ampere (PTPA) method, with voltage limitation imposed by a mathematical model of the machine in steady state. The inductances of the electrical machine are assumed to be independent of the current components Id and Iq, which does not allow for obtaining optimal current components for all types of electrical machines.

[0012] In US patent application 9614473, the current components Id and Iq are determined using the MTPA method with the insertion of a flux-defluxing phase via data tables. The voltage angle is adjusted by a torque controller that uses a desired value and a torque estimator. The use of simple data tables does not always allow for obtaining an optimal control setpoint for the electrical machine. Summary of the invention

[0013] The invention aims to control a rotating electrical machine and to guarantee two objectives: Real-time calculation of optimal current setpoints for all types of electrical machines, and optimal use of the DC bus voltage, enabling improvement of electrical machine efficiency and / or maximum electrical machine performance, even for imperfect optimal current setpoint calculation models.

[0014] To this end, the present invention relates to a method and control system for a synchronous electric machine driven by an inverter, in which the inverter control is determined. The method and control system according to the invention implement a control model that calculates the optimal current setpoints as a function of the torque or speed setpoint and other operating variables, including a corrected operating variable. The correction of this operating variable is implemented by means of a feedback loop controlling the control voltage of the electric machine.

[0015] The invention relates to a method for controlling an electrical machine driven by an inverter equipped with several switching arms. This method involves at least the following steps: a. A torque setpoint is acquired for said electrical machine; b. A setpoint or measurement of a plurality of operating variables of said electrical machine is acquired; c. A correction parameter for one of said operating variables is determined based on an absolute voltage prediction dependent on currents in the phases of said electrical machine and based on voltage setpoints; d. An operating region of said electrical machine is identified based on said torque setpoint and said operating variables of said electrical machine, one of said operating variables being corrected by means of said correction parameter. e.Current setpoints for said electrical machine are determined as a function of said acquired torque setpoint, said acquired operating variables, one of said operating variables being corrected by means of said correction parameter, and by means of a recursive algorithm that solves a system of equations that depends on said identified operating zone and a dynamic model of said electrical machine, said dynamic model of said electrical machine relating in particular the torque of the electrical machine to the currents of said electrical machine; f. A switching control signal for said inverter is determined by means of said current setpoints; and g. Said inverter is controlled by means of said switching control signal.

[0016] According to one embodiment, at least two operating zones of said electrical machine are defined beforehand, delimited by a predefined threshold for at least one operating variable; preferably, said threshold is predefined according to a constraint on the use of said electrical machine, such as a maximum or a minimum of said operating variable considered.

[0017] According to an implementation, said at least one operating variable is chosen from the voltage amplitude of said electrical machine, the current amplitude of said electrical machine, the rotational speed of the rotor of said electrical machine, a temperature of the electrical machine.

[0018] Advantageously, the following operating zones are defined: A first operating zone in which the current amplitude is less than a current threshold, and the voltage amplitude is less than a voltage threshold; a second operating zone in which the current amplitude is greater than or equal to said current threshold, and the voltage amplitude is less than said voltage threshold; a third operating zone in which the current amplitude is greater than or equal to said current threshold, the voltage amplitude is greater than or equal to said voltage threshold, and the rotor speed is less than the maximum power maintenance speed of said electrical machine; a fourth operating zone in which the current amplitude is less than said current threshold, and the voltage amplitude is greater than or equal to said voltage threshold.and possibly a fifth operating zone for which the current amplitude is greater than or equal to said current threshold, the voltage amplitude is greater than or equal to said voltage threshold, and the rotor speed is greater than said maximum power maintenance speed of said electrical machine.

[0019] According to one aspect, the said dynamic model of the said electrical machine is constructed from the di d dt = − R s L d i d + L q L d ωi q + 1 L d ν d di q dt = − R s L q i q − L d L q ωi d + 1 L q ν q − ϕ L q ω C em = 3 2 P ϕ d i q − ϕ q i d the following equations: and with id, iq: the direct and quadrature components of the stator current of said electrical machine, vd, vq: the direct and quadrature components of the voltage of said stator of said electrical machine, ϕ the flux of the rotor magnets of said electrical machine with ϕ the direct flux, and ϕ q the quadrature flux, w the electric angular frequency, R s the stator resistance, L d the direct inductance of said electrical machine, L q the quadrature inductance of said electrical machine, C em the torque of the electrical machine, P is the number of pole pairs of said electrical machine, preferably said stator resistance and said flux of the rotor magnets of said electrical machine are dependent on their respective temperatures and / or preferably the direct inductance Ld and quadrature inductance Lq are non-constant.

[0020] According to one embodiment, said recursive algorithm implements, starting from an appropriate initial or previous value, a recursive solution of Langrange equations of said X y , k = X y , k − 1 − J y , k − 1 ∇ L y , k system of equations, and is written: with X a vector containing the optimal current setpoints and the factors of the constraints of the Lagrange equations of the operating zones, k the time increment, the index y represents said operating zones, ∇ L y , k -1 the gradient of the Lagrange equations, J y,k-1 the Jacobian matrix.

[0021] Advantageously, said recursive algorithm implements a first-order Newton-Raphson method to expand said Jacobian matrix of said system of equations.

[0022] According to one embodiment, said absolute voltage prediction is determined by means of said current setpoints or by means of measurements of said currents in the phases of said electrical machine.

[0023] According to one implementation, said voltage setpoints are determined by means of at least one proportional-integral regulator and said current setpoints.

[0024] Advantageously, the correction parameter is determined by implementing the following steps: i) An absolute voltage prediction is determined; ii) The voltage setpoints are determined; iii) An absolute setpoint voltage is calculated using the formula: V abs sp = V d sp 2 + V q sp 2 , V d sp< and V q sp< being respectively the forward and quadrature voltage setpoints iv) We compare said absolute voltage prediction to said absolute setpoint voltage; and v) We deduce said correction parameter from said comparison.

[0025] Preferably, said control signal is determined by means of said voltage setpoints and vector control.

[0026] According to one aspect, said corrected operating variable is the rotational speed of said electrical machine, the DC bus voltage of the inverter or the standard of the maximum control voltage of the inverter.

[0027] Furthermore, the invention relates to a control system for a synchronous electrical machine comprising an inverter equipped with switching arms, a computer and a memory configured to implement the steps of the control process according to one of the preceding characteristics to control said inverter, preferably said electrical machine is a permanent magnet assisted synchronous reluctance machine or a permanent magnet synchronous machine.

[0028] Other features and advantages of the process and system according to the invention will become apparent from the following description of non-limiting examples of implementations, with reference to the figures attached and described below. List of figures

[0029] There figure 1 illustrates the control system according to one embodiment of the invention. figure 2 illustrates the steps of the control process according to a first embodiment of the invention. figure 3 illustrates the steps of the control process according to a second embodiment of the invention. figure 4 illustrates the steps involved in determining current setpoints according to an implementation of the invention. figure 5 illustrates the steps of the control process according to a third embodiment of the invention. figure 6 illustrates the steps of the control process according to a fourth embodiment of the invention. figure 7 illustrates the steps of the control process according to a fifth embodiment of the invention. figure 8 illustrates several operating zones implemented in an embodiment of the control process according to the invention. Description of the implementation methods

[0030] On the figure 1A schematic and non-limiting installation comprising a rotating electrical machine (MEL) associated, for its control, with a control system (COM) according to the invention (the COM control system implementing the control method according to the invention) is shown. The installation also includes a DC power source, such as a DC voltage bus.

[0031] The MEL rotating electrical machine is a rotating machine with several phases, preferably three phases (alternatively, the electrical machine may have a number of phases that is a multiple of three, for example, six, nine, or twelve, or even four or five phases). Advantageously, the MEL rotating electrical machine may be a synchronous electrical machine. Preferably, the MEL synchronous electrical machine may be a permanent magnet synchronous or synchronous-reluctant electrical machine, in particular a three-phase synchronous-reluctant rotating machine assisted by permanent magnets. Indeed, such electrical machines possess significant flux-shedding capabilities.

[0032] For the illustrated example (which is not exhaustive), the MEL synchronous electric machine has three inputs. Each input corresponds to one phase of a stator (not shown) of the MEL rotating electric machine.

[0033] The COM control system is designed to control the power supply to the rotating machine MEL over time, based on target values ​​and / or measured values ​​of predetermined quantities. These target values ​​and / or measured values ​​are referred to in the following description as operating variables of the electrical machine. For the method and system according to the invention, the COM control system takes into account a torque setpoint Cem* (which can conventionally be derived from a user request for the electrical machine; alternatively, this torque setpoint can be calculated by speed control), and at least one other operating variable VAF, which can be either a measurement or a setpoint.An operating variable of an electrical machine is a quantity that characterizes its operation. This can include electrical variables such as voltage, current, or power, or mechanical variables such as the position, speed, or acceleration of the rotor. Examples of operating variables include the electrical rotational speed ωe of the rotor (not shown) (alternatively, this variable could be the magnitude of the magnetic flux in the phases), the voltage amplitude, the current amplitude, and the temperature. Recall that the electrical rotational speed ωe of the rotor is calculated by multiplying the mechanical rotational speed ωm of the rotor by the number of pole pairs in the synchronous electric machine (MEL).In other words, we can write: ω e =p.ω m with p the number of pole pairs of the synchronous electric machine MEL.

[0034] The COM control system includes an OND inverter and a CAL computer. The COM control system may, if required, include a CAP sensor for angular position and / or angular rotation speed of the MEL electric machine's rotor. Such a sensor allows the angular rotation speed ωm of the electric machine to be determined either directly (in the case of a speed sensor) or by derivation (in the case of a position sensor). Furthermore, the COM control system may include means for measuring the currents (not shown) in the phases of the electric machine, for example, current sensors.

[0035] The OND inverter is configured to route electrical power between the DC source and the synchronous electric machine (MEL). More specifically, the OND inverter is configured to route electrical power between the DC source and each phase of the MEL's stator. The DC source (e.g., a battery) provides a DC bus voltage, denoted V DC.

[0036] The OND inverter has a first input connected to the DC source, and three outputs, each connected to a corresponding phase of the stator of the synchronous electric machine MEL. The OND inverter also has a second input electrically connected to an output of the CAL computer, so that the OND inverter is configured to route electrical power between the DC source and the synchronous electric machine MEL according to a switching control signal applied by the CAL computer to the second input of the OND inverter.

[0037] Preferably, the switching control signal can be such that the OND inverter supplies electrical energy from the DC source to the synchronous electric machine (MEL) so that the MEL operates in a mode commonly referred to as a "motor" and / or a mode commonly referred to as a "generator." Typically, the OND inverter has several switching arms (not shown), preferably at least one switching arm for each phase of the electric machine, to convert the direct current signal from the DC source into an alternating current signal for the phases of the MEL. Each switching arm has at least one controlled switch. Typically, each switch in the switching arms can be controlled by means of pulse-width modulation, known as PWM.

[0038] The optional angular position or angular speed sensor CAP can be configured to measure the angular mechanical position of a rotor (not shown) of the synchronous electric machine relative to the stator (not shown), and to deliver an angular position signal (respectively angular speed) representative of the measured value of the angular position (respectively angular speed) of the rotor of the synchronous electric machine.

[0039] In addition, the optional CAP angular position or angular rotation speed sensor can be configured to apply the angular position signal (respectively angular rotation speed) to a corresponding input of the CAL computer.

[0040] The method for controlling a synchronous electrical machine according to the invention comprises the following steps in real time: Receiving the torque setpoint from the synchronous electric machine, Acquiring a setpoint or a measurement of a plurality of operating variables of the electric machine, Determining a correction parameter for an operating variable, based on a prediction (estimate) of absolute voltage dependent on the currents in the phases of the electric machine or on current setpoints, and based on voltage setpoints (in particular based on the absolute setpoint voltage), Identifying an operating zone of the electric machine based on the torque setpoint and the operating variables of the electric machine, among which one of the operating variables is corrected using the correction parameter, Determining current setpoints for the synchronous electric machine, by means of a recursive algorithm that solves a system of equations based on the identified operating zone,and by means of at least one dynamic model of the electrical machine, the dynamic model of the electrical machine relates: o The torque setpoint, o The operating variable corrected by means of said correction parameter, o The other operating variables (uncorrected), and o The current setpoints: the so-called "direct" and "quadrature" current setpoints, or similarly, the setpoint for the current magnitude and the flux deflux angle) Determination of a switching control signal for the inverter by means of the current setpoints, and Control of the inverter by means of the switching control signal.

[0041] These steps can be implemented in particular by a computer and computer memory, dedicated for real-time processing.

[0042] In the remainder of the description and in the claims, the term "current setpoints" means either "direct" and "quadrature" current setpoints, or current norm and deflux angle, or any other equivalent representation of currents.

[0043] Thus, the control method according to the invention determines, in real time, voltage setpoints based on optimal current setpoints, taking into account an adapted value of the corrected operating variable, as well as an adapted operating range. This adapted value of the corrected operating variable is obtained, in particular, by means of current and torque setpoints or measurements, forming a feedback loop in the control method. Furthermore, the control method does not directly modify the determined voltage setpoints, but only the current setpoints, in order to control said voltage setpoints. In this way, the control method according to the invention can make optimal and robust use of the DC bus voltage in real time.Thanks to this, the control method is robust, for example, robust to variations in the DC bus voltage limit, and the performance and / or efficiency of the synchronous electric machine can be improved, even for an imperfect dynamic model. Furthermore, the control method according to the invention allows for a flux-defluxing operation of the synchronous electric machine. Moreover, these steps can be adapted to all types of rotating synchronous electric machines.

[0044] According to one embodiment of the invention, the following steps can be implemented beforehand and offline: We construct a dynamic model of the electric machine, the dynamic model linking the torque of the electric machine to the currents of the electric machine. We define at least two operating zones of the electric machine, these zones being delimited by one or more thresholds of at least one of the operating variables of the electric machine.

[0045] A threshold for an operating variable of an electrical machine is a limit on the machine's operation related to that variable. These thresholds define operating constraints for the electrical machine. Such a threshold can be defined based on usage constraints of the electrical machine, such as a maximum or minimum value for the operating variable in question. For example: when the operating variable is the voltage amplitude of the electric machine, the threshold can be the maximum voltage of the electric machine; when the operating variable is the current amplitude of the electric machine, the threshold can be the maximum current of the electric machine; when the operating variable is the rotational speed of the electric machine's rotor, the threshold can be the input rotational speed in the MTPV optimization zone (Maximum Torque per Volt).

[0046] These thresholds can be defined based on manufacturer data for the electrical machine or on experimentally obtained data. Furthermore, according to one embodiment of the invention, these thresholds can be defined by taking into account other variable data of the electrical machine, and thus these thresholds can vary over time. Alternatively, the thresholds can be constant over time.

[0047] For example, the maximum voltage may depend on the voltage of the battery powering the electric machine, this battery voltage being considered variable and measurable in real time. For example, the following equation can be written: V smax = mV bat with Vsmax the maximum voltage of the electrical machine, Vbatt the battery voltage and m the Pulse Width Modulation (PWM) index used.

[0048] Operating zones of the electrical machine are identified using the threshold(s) defined for the operating variable(s) under consideration, and in particular for the corrected operating variable. For example, if a single threshold is used for a single variable, a first zone can be identified where the operating variable is below the threshold, and a second zone where the operating variable is greater than or equal to the threshold.

[0049] Identifying the operating zone using the corrected operating variable allows the current setpoints to be adapted to the correction applied, which adds robustness to the process according to the invention.

[0050] For the embodiment in which we consider the three operating variables: voltage amplitude, current amplitude and rotor rotation speed, we can define at least four operating zones of the electrical machine: A first operating zone in which the current amplitude is less than a current threshold (e.g., the maximum current), and the voltage amplitude is less than a voltage threshold (e.g., the maximum voltage); A second operating zone in which the current amplitude is greater than or equal to the current threshold (e.g., the maximum current), and the voltage amplitude is less than the voltage threshold (e.g., the maximum voltage); A third operating zone in which the current amplitude is greater than or equal to the current threshold (e.g., the maximum current), the voltage amplitude is greater than or equal to the voltage threshold (e.g., the maximum voltage), and the rotor speed is less than the maximum power maintenance speed of said electrical machine; A fourth operating zone,A possible fifth operating zone where the current amplitude is less than the current threshold and the voltage amplitude is greater than the voltage threshold; a possible fifth operating zone where the current amplitude is greater than or equal to the current threshold (e.g., the optimal current for achieving maximum torque), the voltage amplitude is greater than or equal to the voltage threshold (e.g., the maximum voltage), and the rotor speed is greater than the maximum power-maintaining speed of the electric machine.

[0051] There figure 8This illustrates, schematically and without limitation, the five zones listed above. The figure shows two curves of the electrical machine: the torque C and the current amplitude I as a function of the rotor speed Ω. These graphs also show a dashed curve called MTPA, representing the voltage threshold (maximum voltage), which depends on the load.

[0052] The first zone, called zone 1 (hatched area), corresponds to an area where the current I is less than the maximum current Imax, and the voltage is less than the maximum voltage. The second zone, called zone 2 (curve segment), corresponds to an area where the current I is greater than or equal to the maximum current Imax, and the voltage V is less than the maximum voltage Vmax. The third zone, called zone 3 (curve segment), corresponds to an area where the current I is greater than or equal to the maximum current Imax, and the voltage V is greater than or equal to the maximum voltage Vmax, and the rotational speed Ω is less than the maximum power maintenance rotational speed denoted Ωp,max. The fourth zone, called zone 4 (unhatched area below the curves), corresponds to an area where the voltage V is greater than the maximum voltage Vmax and the current amplitude is less than the current threshold Imax.The fifth zone, called zone 5 (defluxing zone - portion of curve), corresponds to a zone for which the current I is greater than or equal to the optimal current noted loptiqui is the maximum defluxing current, and the voltage V is greater than or equal to the maximum voltage Vmax, and the rotational speed Ω is greater than or equal to the rotational speed of maximum power maintenance Ωp,max.

[0053] We note that these zones make it possible to distinguish the different zones in which the operation of the machine differs due to operating constraints: at least one of the curves of torque, voltage amplitude, or current amplitude behaves differently when passing from one zone to an adjacent zone.

[0054] For the step of identifying the operating zone of the electric machine, the operating variables of the electric machine and the torque setpoint are determined, and the operating zone of the electric machine is deduced from this. In other words, the operating variables of the electric machine and the torque setpoint are compared to the defined thresholds of these operating variables to deduce, in real time, the operating zone (i.e., the current operating zone of the electric machine).

[0055] For example, the figure 8 If at a given moment we determine that the voltage amplitude is less than the voltage threshold, and that the current amplitude is greater than the current threshold, then we consider for that moment that the machine is in the second operating zone, zone 2. And thus, we then apply the system of equations that corresponds to this zone to determine the current setpoints.

[0056] We construct a dynamic model of the electric machine, which relates the torque of the electric machine to the currents of the electric machine. Several dynamic models of the electric machine can be considered.

[0057] According to one embodiment, the dynamic model of the machine can be constructed di d dt = − R s L d i d + L q L d ωi q + 1 L d ν d di q dt = − R s L q i q − L d L q ωi d + 1 L q ν q − ϕ L q ω electrical using the following equations: and C em = 3 2 P ϕ d i q − ϕ q i d with id, iq: the direct and quadrature components of the stator current of said electrical machine, vd, vq: the direct and quadrature components of the stator voltage of said electrical machine, ϕ the flux of the rotor magnets of said electrical machine with ϕd the direct flux, and ϕq the quadrature flux, we the electric angular frequency ( ωe = PΩ), Rs the stator resistance, Ld the forward inductance of the electrical machine, Lq the quadrature inductance of the electrical machine, Cem the torque of the electrical machine, and P the number of pole pairs of the electrical machine. This model allows for the consideration of variations in the inductance of the electrical machine, which can vary, in particular, according to the forward and quadrature currents. Furthermore, such a model allows for the consideration of the machine's design and its thermal evolution, taking into account, in particular, the stator resistance and the flux of the rotor magnets, which can depend on their respective temperatures.

[0058] In these equations, the forward and quadrature inductances can depend on the forward and quadrature currents. This embodiment is particularly well-suited to synchronous reluctance machines assisted by permanent magnets.

[0059] For another implementation of this embodiment, and for the sake of simplicity in modeling for the optimal calculation of the stator current, we can consider that only the inductance Lq depends on the current θiq (Ld is then considered constant and the mutual inductance is neglected). Therefore, the direct and quadrature components of flux can be denoted respectively ϕ d = L did + ϕ And ϕ q = L q ( iq ) iq. For this variant of the implementation, the simplified dynamic model can be written using the equation: C em = 3 2 P ϕi q + L d − L q i q i d i q

[0060] Furthermore, the voltage amplitude can be defined using the following equation: ν mod = ν d 2 + ν q 2 = − ωL q i q i q + R s i d 2 + ω ϕ + L d i d + R s i q 2 with v mod the voltage amplitude, vd the forward voltage, vq the quadrature voltage.

[0061] Thus, using the dynamic model, a system of equations can be formulated for each operating zone.

[0062] The torque setpoint of the electric machine can be determined conventionally from a user request. Alternatively, the torque setpoint can be determined by speed control.

[0063] For broader application depending on the complexity of the electrical machine used, the method according to the invention implements a recursive solution, preferably a recursive solution of two optimization systems with interlaced constraints based on the current, voltage, and speed of the electrical machine, and integrating the variations in inductance as a function of the stator current components. In this sense, the two optimization systems can be obtained by constructing two Lagrange functions incorporating the various constraints considered for this method.

[0064] For the current setpoint determination step, a recursive algorithm with a single computation time step is implemented. A recursive algorithm solves a problem by calculating smaller instance solutions to the same problem. This algorithm is called recursive because the current setpoint at time k (tk in discrete time) depends on the current setpoint at time k-1 (i.e., the previous discrete time tk-1). This algorithm aims to solve the minimization problem arising from the system of equations for the determined operating region, thus enabling an optimal level of performance for the electrical machine. The recursive algorithm preferably uses only the value of the previous time step. This recursive algorithm also has the advantage of not using any data tables or mappings to determine the current setpoints.Furthermore, this algorithm combined with the correction allows for optimal use of the DC bus voltage, even for imperfect optimal current setpoint calculation models.

[0065] Several methods can be implemented for solving systems of equations.

[0066] According to one embodiment of the invention, it is possible to solve the Lagrange equations of the system of equations and to implement a first-order Newton-Raphson method. Thus, one can write: X y , k = X y , k − 1 − J y , k − 1 − 1 ∇ L y , k − 1 with X a vector containing the optimal current setpoints from the Lagrange equations of the operating zones, k the time increment, the index y represents the operating zones, ∇ Ly,k- 1 the gradient of the Lagrange equations at the previous instant k-1 , Jy,k- 1 the Jacobian matrix at the previous instant t k-1 .

[0067] We establish a first current optimization system, called Sys1, during which the torque setpoint can be achieved in the following manner: min i d , i q 1 2 i d 2 + i q 2 C em ∗ = 3 2 P ϕ + L d i q − L q i d i q i d i q i d 2 + i q 2 ≤ I smax − ωL q i d i q i q + R s i d 2 + ω ϕ + L d i q i d + R s i q 2 ≤ V smax

[0068] A second current optimization system, called Sys2, is implemented, during which the torque setpoint cannot be achieved, in the following manner: min i d , i q 1 2 C em ∗ − 3 2 P ϕ + L d − L q i q i d i q 2 i d 2 + i q 2 ≤ I s max − ωL q i q i q + R s i d 2 + ω ϕ + L d i d − R s i q 2 ≤ V s max

[0069] By minimizing the total loss under an equality constraint on the torque and another inequality constraint on the tension, we can write the Lagrangian of the first system Sys1 as: Lag 1 = 1 2 i d 2 + i q 2 + λ Cem C em ∗ − C em + λ ν max V s max − V mod C em = 3 2 P ϕ + L d i d i q − L q i d i q i d i q V mod = − ωL q i d i q i q + R s i d 2 + ω ϕ + L d i d i q i d + R s i q 2

[0070] C em And C em ∗ are, respectively, the electromagnetic couple and its setpoint, V mod is the amplitude of the voltage, λ Cem is the desired electromagnetic torque Lagrange factor and λ vmax is the maximum voltage Lagrange factor of the Sys1 system.

[0071] Then, we can analytically determine the gradient of this Lagrangian: ∇ Lag 1 = ∂ Lag 1 ∂ i d ∂ Lag 1 ∂ i q ∂ Lag 1 ∂ λ Cem ∂ Lag 1 ∂ λ ν max T = dLag 1 dLag 2 dLag 3 dLag 4 T

[0072] The Jacobian matrix can be calculated using finite difference: J Sys 1 = dLag 1 − dLag 1 i d di d dLag 1 − dLag 1 i q di q dLag 1 − dLag 1 λ Cem dλ Cem dLag 1 − dLag 1 λ νmax dλ νmax dLag 2 − dLag 2 i d di d dLag 2 − dLag 2 iq di q . . dLag 3 − dLag 3 i d di d . . . dLag 4 − dLag 4 i d di d . . dLag 4 − dLag 4 λ ν max dλ ν max

[0073] With dLag 1 … 4 i d = dLag 1 … 4 i d + di d , i q , λ Cem , λ ν max dLag 1 … 4 i q = dLag 1 … 4 i d , i q + di q , λ Cem , λ ν max dLag 1 … 4 λ Cem = dLag 1 … 4 i d , i q , λ Cem + dλ Cem , λ ν max dLag 1 … 4 λ ν max = dLag 1 … 4 i d , i q , λ Cem , λ ν max + dλ ν max

[0074] Thus, for the embodiment with four operating zones ( figure 8 ), the current setpoints for zone 1 can be obtained using: X 11 , k = X 11 , k − 1 − J Sys 1 _ 1 − 1 ∇ Lag 11 avec X 11 = i d i q λ Cem T

[0075] And for zone 4, using: X 1 , k = X 1 , k − 1 − J Sys 1 − 1 ∇ Lag 1 avec X 12 = i d i q λ Cem λ ν max T

[0076] With J sys1_1 a (3x3) matrix selected from the matrix J sys1 as a function of the vector X 11.

[0077] By minimizing the difference in the electromagnetic torque under a constraint of inequality on the voltage and another of inequality on the current, we can write the Lagrangian of the second system Sys2 as: Lag 2 = C em ∗ − 3 2 P ϕ + L d i d i q − L q i d i q i d i q + γ ν max V mod − V s max + γ i max I mod − I s max I mod = i d 2 + i q 2 V mod = − ωL q i d i q i q + R s i d 2 + ω ϕ + L d i d i q d + R s i q 2 C em ∗ is the electromagnetic torque setpoint, V mod is the amplitude of the voltage, γ imax is the maximum current Lagrange factor and γ vmax is the maximum voltage Lagrange factor of the Sys2 system.

[0078] Then, we can analytically determine the gradient of this Lagrangian: ∇ Lag 2 = ∂ Lag 2 ∂ i d ∂ Lag 2 ∂ i q ∂ Lag 2 ∂ γ ν max ∂ Lag 2 ∂ γ i max T = dLag 21 dLag 22 dLag 23 dLag 24 T

[0079] The Jacobian matrix can be calculated using finite difference: J Sys 2 = dLag 21 − dLag 21 i d di d dLag 21 − dLag 21 i q di q dLag 21 − dLag 21 γ ν max dγ ν max dLag 21 − dLag 21 γ i max dγ i max dLag 22 − dLag 22 i d di d dLag 22 − dLag 22 iq di q ⋅ ⋅ dLag 23 − dLag 23 i d di d ⋅ ⋅ ⋅ dLag 24 − dLag 24 i d di d ⋅ ⋅ dLag 24 − dLag 24 γ i max dγ i max

[0080] With dLag 21 … 4 i d = dLag 21 … 4 i d + di d , i q , γ ν max , γ i max dLag 21 … 4 i q = dLag 21 … 4 i d , i q + di q , γ ν max , γ i max dLag 21 … 4 λ ν max = dLag 21 … 4 i d , i q , γ ν max + dγ ν max , γ i max dLag 21 … 4 λ i max = dLag 21 … 4 i d , i q , γ ν max , γ i max + dγ i max

[0081] Thus, for the embodiment with four operating zones ( figure 8 ), the current setpoints for zone 2 can be obtained using: X 21 , k = X 21 , k − 1 − J Sys 2 _ 1 − 1 ∇ Lag 21 avec X 21 = i d i q γ i max T

[0082] For zone 3, using: X 2 , k = X 2 , k − 1 − J Sys 2 − 1 ∇ Lag 2 avec X 2 = i d i q γ ν max γ i max T

[0083] For zone 5, using: X 2 , k = X 2 , k − 1 − J Sys 2 _ 2 − 1 ∇ Lag 22 avec X 22 = i d i q γ ν max T

[0084] With J sys2_1 and J sys2_2 selected (3x3) matrices of matrix J sys2 as a function of vectors X 21 and X 22.

[0085] There figure 4This illustrates, schematically and without limitation, the steps for determining current setpoints according to an embodiment of the invention. First, a torque setpoint for the electric machine, Cem*, is determined in real time, and at least one operating variable, VAF, of the electric machine is acquired. Then, using operating zones identified by means of defined thresholds, SEU, an operating zone, ZON, of the electric machine is determined in real time, as a function of the torque setpoint, Cem, and at least one operating variable, VAF. Prior to this, a dynamic model, MOD, of the electric machine has been constructed. In real time, a recursive algorithm, ALR, is implemented which uses the dynamic model, MOD, and the identified operating zone, ZON, to determine current setpoints, id and iq.The recursive ALR algorithm depends in particular on the torque setpoint, the operating variables, including the corrected operating variable. The real-time steps that allow the determination of the current setpoints id, iq are denoted DET (and correspond to the DET step of the . figures 2 , 3 , And 5 à 7 which are described below).

[0086] According to one embodiment of the invention, the absolute voltage prediction can be determined using current setpoints by an estimator. Alternatively, the absolute voltage prediction can be determined using phase current measurements by an estimator.

[0087] Furthermore, the estimator implemented in these alternatives can depend on the corrected value of the operating variable using the correction parameter, or on both the correction parameter and the operating variable. Thus, the estimator can accurately reflect the various steps involved in determining the inverter control.

[0088] According to one example implementation, the estimator can be constructed using at least one nonlinear flow map.

[0089] Conventionally, voltage setpoints can be determined using at least one proportional-integral (PI) controller, referred to hereafter as a PI controller. Such a PI controller generates the voltage setpoints based on the current setpoints. In one embodiment, the control method may implement a first PI controller for the "forward" voltage setpoint and a second PI controller for the "quadrature" voltage setpoint. In this embodiment, the PI controller may also take into account the currents measured in the phases of the electrical machine.

[0090] Alternatively, any similar method can be implemented for determining voltage setpoints.

[0091] According to one aspect of the invention, the correction parameter can be determined by implementing the following steps: The prediction of absolute voltage, denoted V abs pred ,For example, using the model employed in the recursive algorithm, the voltage setpoints are determined, with Vdspi and Vqspi being the forward and quadrature voltage setpoints respectively. An absolute setpoint voltage is calculated, for example, using the formula V abs sp = V d sp 2 + V q sp 2 or by means of any similar formula, the prediction of absolute tension is compared V abs pred and said absolute setpoint voltage V abs sp , And we deduce the correction parameter from the comparison.

[0092] Then, one of the operating variables is corrected using the correction parameter.

[0093] According to one option of the invention, the correction parameter can be a multiplicative parameter (a gain), an additive parameter (an offset, also called a deviation), or a difference between the predicted absolute voltage value and the absolute setpoint voltage. When the correction parameter is a gain, then the corrected control value can be the product of the operating variable and the gain. When the correction parameter is a corrected value of an electrical machine variable, then the corrected control value can be obtained by any mathematical operation between the operating variable and the corrected value of a machine variable.

[0094] According to one embodiment, the calculation of the correction parameter Pco can be carried out by a control stage, for example of the Proportional Integral type.

[0095] According to one embodiment of the invention, the inverter switching control signal can be determined by means of voltage setpoints and vector control, in particular conventional Space Vector Modulation (SVM) control, or any similar method. The determined control signal corresponds to the control signals sent to the switching arms of the inverter, which convert the DC bus signal into AC signals.

[0096] Thus, the inverter switching control signal can be determined using the voltage setpoints by means of the following two steps: Determination of voltage setpoints from current setpoints, for example by means of at least one PI regulator, and Determination of switching control signals from voltage setpoints, for example by means of an SVM type vector control.

[0097] For the embodiment in which an operating variable of the electric machine is the electrical rotational speed, this speed can be estimated or measured. In the case of measurement, this can be achieved by measuring the mechanical rotational speed of the rotor using a sensor (in particular, a position or angular velocity sensor of the rotor) and multiplying the measured mechanical rotational speed by the number of pole pairs of the electric machine. Alternatively, the electrical position can be determined directly by a rotation sensor. In the case of estimating the electrical rotational speed, any of the conventional sensorless estimation methods can be implemented.

[0098] The control system according to the invention comprises an inverter, optionally a position and / or speed sensor for the rotor of the electric machine, and a computer and memory configured to implement the steps of the control process according to any of the variants or combinations of variants described above. Furthermore, the control system may include means for measuring the currents (not shown) in the phases of the electric machine, for example, current sensors.

[0099] According to one embodiment of the invention, the rotating electrical machine can be a synchronous electrical machine, preferably a synchronous-reluctance electrical machine assisted by permanent magnets. Indeed, the method and system according to the invention are particularly well-suited to this type of electrical machine, notably because the invention allows for consideration of the constraints and operation of all types of machines.

[0100] There figure 2 illustrates, schematically and without limitation, the steps of the process according to a first embodiment of the invention. The elements identical to the figure 1 are not described a second time. Therefore, only the CAL calculator is detailed.

[0101] The control process implemented by the CAL computer initially involves a step to determine the current setpoints id sp< and iq sp< from the torque setpoint Cem* and operating variables VAF, including the corrected operating variable VAR corr<. In the figure, only one arrow is shown for the operating variables VAF for the sake of simplicity. However, this arrow actually represents a plurality of operating variables VAF. This current setpoint determination step is implemented using at least one recursive DET algorithm (notably as illustrated in figure 4 ).

[0102] At least one PI regulator (preferably two PI regulators or any similar method) is implemented to determine the voltage setpoints Vdsp< and Vqsp< from the current setpoints idsp and iqsp.

[0103] The voltage setpoints Vdsp< and Vqsp< are converted into a switching control signal for the OND inverter by means of a spatial vector control SVM, or any similar method.

[0104] Furthermore, the control process includes a COR step for determining the correction parameter Pco. This correction is implemented based on an absolute setpoint voltage. V abs sp which is obtained directly from the voltage setpoints Vdsp< and Vqsp< and from a predicted absolute voltage V abs pred which is obtained by a PRE prediction (or estimation) step based on the voltage setpoints id sp< and iq sp< . As illustrated, the PRE prediction step can also take into account the corrected operating variable VAR corr< .

[0105] This correction factor Pco corrects an input operating variable VAR of the control process, so as to determine the corrected operating variable VAR corr< used in the control model MCO, thus forming a feedback loop.

[0106] There figure 3 illustrates, schematically and without limitation, the steps of the process according to a second embodiment of the invention. The elements identical to the figure 2 are not described a second time.

[0107] For this embodiment, the predicted absolute voltage V abs pred is obtained by a PRE (or estimation) prediction step based on the voltage measurements idmes and iqmes. The voltage measurements idmes and iqmes are obtained by applying a Park transformation to the currents ia, ib, ic measured by a current sensor in the phases denoted a, b, and c of the electrical machine. Such a Park transformation can be performed by applying a relationship of the type: i d i q = cos θ e cos θ e − 2 π 3 cos θ e + 2 π 3 − sin θ e − sin θ e − 2 π 3 − sin θ e + 2 π 3 i A i B i C

[0108] With θe the electrical angular position of the rotor relative to the stator of the electric machine (as a reminder, the electrical angular position corresponds to the multiplication of the mechanical angular position of the rotor relative to the stator of the electric machine and the number of pole pairs of the electric machine).

[0109] There figure 5illustrates, schematically and without limitation, the steps of the process according to a third embodiment of the invention. The elements identical to the embodiment of the figure 2 are not described a second time.

[0110] For this embodiment, the corrected operating variable is the magnitude of the maximum control voltage Vmax, thus forming a corrected value Vmax corr<. This magnitude of the maximum control voltage Vmax can be defined by: V max = V dc × M with Vdc the DC bus voltage and M a modulation index.

[0111] Furthermore, for this embodiment, the PRE prediction can also or alternatively depend on the correction parameter Pco.

[0112] This embodiment can also be associated with the embodiment of the figure 3 .

[0113] There figure 6illustrates, schematically and without limitation, the steps of the process according to a fourth embodiment of the invention. The elements identical to the embodiment of the figure 2 are not described a second time.

[0114] For this embodiment, the corrected operating variable is the DC bus voltage Vdc, thus forming a corrected value Vdc corr< .

[0115] Furthermore, for this embodiment, the PRE prediction can also or alternatively depend on the correction parameter Pco.

[0116] This embodiment can also be associated with the embodiment of the figure 3 .

[0117] There figure 7 illustrates, schematically and without limitation, the steps of the process according to a fifth embodiment of the invention. The elements identical to the embodiment of the figure 2 are not described a second time.

[0118] For this embodiment, the corrected operating variable is the electrical rotational speed ωe of the electric machine rotor, thus forming a corrected value ωe corr< .

[0119] Furthermore, for this embodiment, the PRE prediction can also or alternatively depend on the correction parameter Pco.

[0120] This embodiment can also be associated with the embodiment of the figure 3 .

Claims

1. Method for controlling an electric machine (MEL) controlled by an inverter (OND) provided with several switching arms, wherein the following steps are implemented: a. a torque setpoint (Cem*) of said electric machine (MEL) is acquired; b. a setpoint or a measurement of a plurality of operating variables (VAF) of said electric machine (MEL) is acquired; said method being characterized in that at least the following steps are implemented: c. a correction parameter (Pco) for one of said operating variables (VAF) is determined as a function of an absolute voltage prediction which is dependent on currents in the phases of said electric machine and as a function of voltage setpoints; d. a zone of operation (ZON) of said electric machine (MEL) is identified as a function of said torque setpoint and said operating variables of said electric machine, one of said operating variables being corrected by means of said correction parameter; e. current setpoints of said electric machine are determined (DET) as a function of said acquired torque setpoint (Cem*), said acquired operating variables (VAF), one of said operating variables being corrected (VARcorr) by means of said correction parameter (Pco), and by means of a recursive algorithm (ALR) that solves a system of equations that depends on said identified zone of operation and on a dynamic model (MOD) of said electric machine, said dynamic model (MOD) of said electric machine relating in particular the torque of the electric machine to the currents of said electric machine; f. a switching control signal for said inverter (OND) is determined by means of said current setpoints; and g. said inverter (OND) is controlled by means of said switching control signal.

2. Control method according to Claim 1, wherein at least two zones of operation (ZON) of said electric machine, which are delimited by a predefined threshold (SEU), are defined beforehand for at least one operating variable (VAF), and said threshold (SEU) is preferably predefined as a function of a constraint of use of said electric machine, such as a maximum or a minimum of said operating variable under consideration.

3. Control method according to either of the preceding claims, wherein said at least one operating variable (VAF) is chosen from the voltage amplitude of said electric machine, the current amplitude of said electric machine, the rotational speed of the rotor of said electric machine, and a temperature of the electric machine.

4. Control method according to Claim 3, wherein the following zones of operation are defined: • A first zone of operation (zone1) for which the current amplitude is lower than a current threshold and the voltage amplitude is lower than a voltage threshold, • A second zone of operation (zone2) for which the current amplitude is greater than or equal to said current threshold and the voltage amplitude is lower than said voltage threshold, • A third zone of operation (zone3) for which the current amplitude is greater than or equal to said current threshold, the voltage amplitude is greater than or equal to said voltage threshold, and the rotational speed of the rotor is lower than the maximum power-maintaining rotational speed of said electric machine, • A fourth zone of operation (zone4) for which the current amplitude is lower than said current threshold and the voltage amplitude is greater than or equal to said voltage threshold, and • Possibly a fifth zone of operation (zone5) for which the current amplitude is greater than or equal to said current threshold, the voltage amplitude is greater than or equal to said voltage threshold, and the rotational speed of the rotor is greater than said maximum power-maintaining rotational speed of said electric machine.

5. Control method according to one of the preceding claims, wherein said dynamic model (MOD) of said electric machine is constructed based on the following equations: di d dt = − R s L d i d + L q L d ωi q + 1 L d ν d di q dt = − R s L q i q − L d L q ωi d + 1 L q ν q − ϕ L q ω and C em = 3 2 P ϕ d i q − ϕ q i d with id, iq: the direct and quadrature components of the stator current of said electric machine, vd, vq; the direct and quadrature components of the voltage of said stator of said electric machine, Φ the flux of the magnets of the rotor of said electric machine, with Φd the direct flux, and Φq the quadrature flux, ω the electric angular frequency, Rs the stator resistance, Ld the direct inductance of said electric machine, Lq the quadrature inductance of said electric machine, Cem the torque of the electric machine, P is the number of pole pairs of said electric machine, said stator resistance and said flux of the magnets of the rotor of said electric machine preferably being dependent on their respective temperatures and / or the direct inductance Ld and quadrature inductance Lq preferably being non-constant.

6. Control method according to one of the preceding claims, wherein said recursive algorithm (ALR), starting from an appropriate initial or previous value, recursively solves Lagrange equations of said system of equations, and is written: Xy,k = Xy,k-1 (Jy,k)-1∇Ly,k with X a vector containing the optimal current setpoints and the factors of the constraints of the Lagrange equations of the zones of operation, k the time increment, the index y represents said zones of operation, ∇Ly,k-1 the gradient of the Lagrange equations, Jy,k-1 the Jacobian matrix.

7. Control method according to Claim 6, wherein said recursive algorithm (ALR) implements a first-order Newton-Raphson method to develop said Jacobian matrix of said system of equations.

8. Control method according to one of the preceding claims, wherein said absolute voltage prediction is determined by an estimator by means of said current setpoints (idsp, iqsp) or by means of measurements of said currents (idmes, iqmes) in the phases of said electric machine (MEL).

9. Control method according to one of the preceding claims, wherein said voltage setpoints (Vdsp, Vqsp) are determined by means of at least one proportional-integral controller (PI) and said current setpoints (idsp, iqsp).

10. Control method according to either of Claims 8 and 9, wherein said correction parameter (Pco) is determined by implementing the following steps: i) an absolute voltage prediction (vabspred) is determined; ii) said voltage setpoints (Vdsp, Vqsp) are determined; iii) a setpoint absolute voltage is calculated using the formula: V abs sp = V d sp 2 + V q sp 2 , where Vdsp and Vqsp are the direct and quadrature voltage setpoints, respectively; iv) said absolute voltage prediction (vabspred) is compared with said setpoint absolute voltage (vabssp); and v) said correction parameter (Pco) is deduced therefrom on the basis of said comparison.

11. Control method according to either of Claims 9 and 10, wherein said control signal is determined by means of said voltage setpoints (Vdsp, Vqsp) and vector control (SVM) .

12. Control method according to one of the preceding claims, wherein said corrected operating variable (VARcorr) is the rotational speed of said electric machine, the DC bus voltage of the inverter or the norm of the maximum control voltage of the inverter.

13. System for controlling a synchronous electric machine (MEL) comprising an inverter (OND) provided with switching arms, a computer (CAL) and a memory which is configured to implement the steps of the control method according to one of the preceding claims in order to control said inverter (OND), said electric machine (MEL) preferably being a permanent-magnet-assisted synchronous reluctance machine or a permanent-magnet synchronous machine.