Method for controlling an electrical machine with voltage setpoint correction

Abstract

The present invention relates to a method for controlling an electrical machine by means of pulse width modulation from a voltage setpoint. For this method, the voltage setpoint is corrected when the voltage setpoint is partially saturated during a modulation period. Then, the switching setpoint is determined at the corrected voltage setpoint, this switching setpoint then being applied to the switches of the inverter arms. Figure 1 to be published

Description

Title of the invention: Method for controlling an electrical machine with correction of the voltage setpoint Technical field

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

[0002] It is known to use rotating electrical machines such as permanent magnet synchronous electrical machines or synchro-reluctant machines, in particular synchro-reluctant machines assisted by permanent magnets. Such electrical machines are, for example, used in the field of propulsion, for example for generating engine torques on board a vehicle such as a motor vehicle. They can also be used in stationary applications.

[0003] Document FR 3 051 296 A1 describes, for example, a synchro-reluctant machine assisted by permanent magnets.

[0004] The maximum power of three-phase motors is limited by the current constraint and the voltage constraint. For electrical machines controlled by a voltage source inverter, the current constraint is given by the maximum current that can flow in the phases of the inverter or the motor, while the voltage constraint is given by the DC supply voltage of the inverter. To maximize the achievable power, the inverter must fully exploit the available DC voltage and thus allow the application of maximum phase voltage commands.

[0005] Figure 1 illustrates, schematically and in a non-limiting manner, a power supply system for an electrical machine (not shown) comprising three phases A, B and C. The power supply system comprises an inverter OND comprising three switching arms BC, each arm comprising two switches (AHS, ALS, BHS, BLS, CHS, CLS) and an output (A, B and C), each output being crossed by a current ia, ib, ic. The inverter OND is powered by a direct voltage source DC, of ​​value noted Vin. Capacitors are placed in parallel with the switching arms, each with a direct voltage of vdc. 2

[0006] If we consider the voltage source inverter, the applied phase voltage has the form of rectangular waves with peak-to-peak amplitude equal to VDc-

[0007] However, operating the inverter at or near the maximum control voltage has a number of drawbacks, such as increasing the harmonic distortion rate (THD) of the currents, or increasing the ripple torque. (from the English "ripple"), and also impacts the dynamics of the system in load transients. In addition, at high speed, conventional control algorithms generally fail to properly control the applied voltage, due to the appearance of low-frequency harmonics that can make the control unstable.

[0008] Ideally, the voltages applied to a three-phase electrical machine would be three sinusoidal voltages phase-shifted by 0°, 120° and 240° respectively. However, due to its discrete nature, the voltage inverter can only apply voltages in oscillating square waves.

[0009] Pulse Width Modulation (PWM) is based on the idea that a sinusoid (in this case the control voltage of an electrical machine) can be "cut" into a series of constant values. Indeed, the PWM method is the classic method for obtaining a sinusoidal voltage at the output of the inverter, for a given DC voltage at the input of the inverter. The idea is to apply the desired voltage setpoint on average over a predefined time period called the cutting period, from pulses of DC voltage amplitude but of width proportional to the desired voltage setpoint.

[0010] By this PWM method, the reference voltage is updated at each modulation period and takes the value of the ideal desired control voltage in the middle of the modulation period. As long as the voltage setpoint values ​​are between de and jjV (corresponding respectively to the minimum saturation voltage and to 2 2 the maximum saturation voltage), the voltage setpoint can be applied on average at the inverter output. This operating zone is called the linear operating zone of the inverter.

[0011] Figure 2 illustrates the ratio Va / Vde of the voltage of a phase a Va by the voltage of the continuous source Vdc with respect to the electrical angle of the electrical machine 0, in °. On the ordinate, the values ​​(not indicated) of -0.5 and +0.5 correspond to the saturation voltages (in fact, this then corresponds respectively to ~vdc and On this 2 2 graph, we represent the curve of the voltage setpoint Vref and the discretized voltage VdiS by the PWM process for phase a of the electric machine. The modulation period is noted TPWM, the midpoint of this period is indicated by the point P. For the example illustrated, the curve of the voltage setpoint Vref remains lower than 0.5: there is therefore no saturation over this modulation period. Over a TPWM modulation period, the ideal average of the voltage setpoint corresponds to the voltage setpoint when the hatched areas '+' and '-' are the same. The approximation is acceptable if the ideal voltage setpoint is close to a straight line (which is favored over a short TPWM modulation period). vdc: , vdc , the control voltage is saturated at vdc or , ix . This area of ​​? ' 2 2 7 operation is called non-linear operating zone, or overmodulation zone, and thus causes two types of errors on the applied voltage: 1. An error due to saturation: the fundamental of the applied control voltage is reduced. 2. An error due to the time discretization linked to the modulation. Since the saturation of the control creates a discontinuity in the normally observed quasi-linear trajectory, the voltage setpoint can no longer be correctly approximated by a straight line, and therefore a voltage error is generated. In addition, this error depends on the start time of the TPWM modulation period relative to the ideal control. As a result, the general shape of the voltage setpoint Vref is likely to change from one electrical Telec period to the next, which then induces low-frequency harmonics on the control voltage.

[0013] Figure 3 is a graph similar to Figure 2 for which the voltage setpoint Vref exceeds the saturation value (0.5) near the midpoint P. In this figure, the ideal voltage Empty to obtain the voltage setpoint is also illustrated. This situation creates a saturation of the voltage setpoint and a voltage error.

[0014] Figure 4 is a graph similar to Figure 2 for which the voltage setpoint Vref exceeds the saturation value (0.5) at the beginning of the modulation period. In this figure, the ideal voltage Empty is also illustrated to obtain the voltage setpoint. This situation generates a saturation of the voltage setpoint but does not create any voltage error due to discretization. However, an error due to saturation is generated.

[0015] The problem is even more acute if the voltage setpoint Vref and the ideal voltage (i.e. the voltage if there were no saturation) are completely saturated; the applied voltage is then a square wave, known as a "full wave". In this situation, a slight difference in phasing between the sampling instant and the instant of switching from one saturation to the other of Vref creates a voltage error that changes sign. Figures 5 and 6 are graphs similar to the figure illustrating this situation in two cases. For [Fig.5], the switch is made before the midpoint P, while for [Fig.6], the switch is made after the midpoint P. These situations then create voltage errors illustrated by the hatched areas "+" and "-". Prior art

[0016] To resolve these different problems, several technical solutions have been proposed.

[0017] For example, patent application US2012 / 0169263 describes a method for adjusting the duty cycle of a pulse width modulation method of a space vector modulation (SVM) method, with the aim of reducing errors in the voltage control signals, thereby reducing torque and current oscillations. However, the implementation of this method is limited to the SVM method, which notably involves common compensation for all phases of the electrical machine. Furthermore, the adjustment of the duty cycle depends on unknown constants, which makes this method difficult to implement.

[0018] Furthermore, patent application US2019 / 0097562 describes a pulse width modulation control method, the purpose of which is to smooth the transitions between the overmodulation mode and the PWM mode, in particular during high-speed operations of the electrical machine. However, the method described in this application requires numerous switching operations. These numerous switching operations generate significant losses. Summary of the invention

[0019] The aim of the invention is to control an electrical machine in a simple manner, by limiting voltage errors, harmonics, current and torque ripples, as well as switching losses. For this purpose, the present invention relates to a method for controlling an electrical machine by means of pulse width modulation from a voltage setpoint. For this method, the voltage setpoint is corrected when the voltage setpoint is partially saturated during a modulation period. Then, the switching setpoint is determined at the corrected voltage setpoint, this switching setpoint then being applied to the switches of the inverter arms. The correction of the voltage setpoint makes it possible to reduce voltage errors in the overmodulation zone and in the full-wave operating zone of the electrical machine.

[0020] Furthermore, the invention relates to a control system implementing such a control method.

[0021] The invention relates to a method for controlling an electrical machine by means of pulse width modulation from a voltage setpoint of the phases of said electrical machine, said electrical machine being controlled by an inverter provided with a plurality of switching arms and powered by a DC voltage source. For this method, the following steps are implemented: a. At least one modulation period is detected for which said voltage setpoint is not entirely saturated at a single voltage by the voltage of

[0022]

[0023]

[0024]

[0025]

[0026]

[0027] the continuous voltage source over said modulation period, and the instant of variation of the saturation of said voltage setpoint is detected; b. For each detected modulation period, a correction of said voltage setpoint is applied, such that the corrected voltage setpoint is the weighted average of the voltage setpoint of two time intervals separated by said instant of variation of the saturation over the modulation period; c. A switching signal is determined for each switching arm of said inverter as a function of said corrected voltage setpoint; and d. Each switching arm of said inverter is controlled by applying said determined switching instruction. According to one embodiment, for each detected modulation period, and when said voltage setpoint changes from a first saturated state to a second saturated state or when the voltage setpoint is overmodulated, a switching setpoint is determined, by carrying out the switching interval at one end of the modulation period. Advantageously, the switching interval begins at the beginning of the modulation period when the voltage setpoint varies between maximum saturation and minimum saturation. Advantageously, the switching interval is terminated at the end of the modulation period when the voltage setpoint varies between minimum saturation and maximum saturation. According to one implementation, when the voltage setpoint varies from an unsaturated voltage to a saturated voltage, said corrected voltage setpoint is determined using the following formula: y _ vnjrw TnuSat with Vrefcor the corrected voltage setpoint, Tpwm the modulation period, Vrefr the saturated voltage setpoint during the saturated time interval TxaJ of the modulation period, VrefT the unsaturated voltage setpoint during the unsaturated time interval TlwSat of the modulation period, the saturated time interval being separated from the unsaturated time interval at said switching variation instant. In one aspect, the unsaturated voltage setpoint is the average of the voltage setpoint voltage during the unsaturated time interval of the modulation period. Alternatively, when the voltage setpoint varies from a first saturated voltage to a second saturated voltage, we determine said corrected voltage setpoint at using the following formula '-y _ V refCor VrefT„, T sat2 with VrefCor la T PWM corrected voltage setpoint, TPWM the modulation period, ^refT^ the saturated voltage setpoint during the first saturation time interval Tsatl of the modulation period, VrefTs, the saturated voltage setpoint during the second saturation time interval TSat2 of the modulation period, the first saturation time interval being separated from the second saturation time interval at said switching variation instant.

[0028] Furthermore, the invention relates to a system for controlling an electrical machine comprising an inverter provided with switching arms, a computer and a memory configured to implement the steps of the control method according to one of the preceding characteristics.

[0029] According to one embodiment, said electrical machine is a synchro-reluctant machine assisted by permanent magnets or a permanent magnet synchronous machine.

[0030] Other characteristics and advantages of the method and system according to the invention will appear on reading the following description of non-limiting examples of embodiments, with reference to the appended figures described below. List of figures

[0031] [Fig.l]

[0032] [Fig.l], already described, illustrates an inverter according to an alternative embodiment.

[0033] [Fig.2]

[0034] [Fig.2], already described, is a graph illustrating the MLI method in a linear area.

[0035] [Fig.3]

[0036] [Fig. 3], already described, is a graph illustrating the MLI method according to the prior art in an overmodulation zone.

[0037] [Fig.4]

[0038] [Fig. 4], already described, is a graph illustrating the MLI method according to the prior art in an overmodulation zone.

[0039] [Fig.5]

[0040] [Fig. 5], already described, is a graph illustrating the MLI method according to the prior art in a “full wave” zone.

[0041] [Fig.6]

[0042] [Fig. 6], already described, is a graph illustrating the MLI method according to the prior art in a “full wave” zone.

[0043] [Fig.7]

[0044] [Fig.7] is a graph illustrating the application of the method according to one embodiment of the invention for the situation of [Fig.3].

[0045] [Fig. 8]

[0046] [Fig.8] is a graph illustrating the application of the method according to one embodiment of the invention for the situation of [Fig.3].

[0047] [Fig.9]

[0048] [Fig.9] is a graph illustrating the application of the method according to one embodiment of the invention for the situation of [Fig.5].

[0049] [Fig. 10]

[0050] [Fig. 10] is a graph illustrating the application of the method according to one embodiment of the invention for the situation of [Fig.5].

[0051] [Fig. 11]

[0052] [Fig. 11] is a graph illustrating the switching signal for the example of [Fig.10],

[0053] [Fig. 12]

[0054] [Fig.12] is an enlargement of [Fig.11].

[0055] [Fig. 13]

[0056] [Fig. 13] is a graph illustrating the corrected switching signal according to one embodiment of the invention.

[0057] [Fig. 14]

[0058] [Fig. 14] illustrates the control of an electrical machine according to one embodiment of the invention. Description of the embodiments

[0059] [Fig. 14] shows, schematically and in a non-limiting manner, an installation comprising a rotating electrical machine MEL associated, for its control, with a COM control system according to the invention (the COM control system implementing the control method according to the invention). The installation also comprises a DC source of electrical energy, such as a direct voltage bus.

[0060] The rotating electrical machine MEL is a rotating machine with several phases, preferably three phases (alternatively the electrical machine may comprise a number of phases that is a multiple of three, for example six, nine or twelve, or else four or five phases). Advantageously, the rotating electrical machine MEL may be a synchronous electrical machine. Preferably, the synchronous electrical machine MEL may be a permanent magnet or synchro-reluctant synchronous electrical machine, in particular a three-phase synchro-reluctant rotating machine assisted by permanent magnets.

[0061] For the illustrated example (non-limiting), the synchronous electric machine MEL has three inputs. Each input corresponds to a phase of a stator (not re- presented) of the rotating electric machine MEL.

[0062] The control system COM is intended to control, over time, the power supply of the rotating machine MEL as a function of target values ​​and / or measured values ​​of predetermined quantities. These target values ​​and / or measured values ​​are referred to in the remainder of the description as operating variables of the electrical machine. For the illustrated embodiment, the control system COM can take into account a torque setpoint Cem* (which can conventionally come from a request from the user of the electrical machine, alternatively this torque setpoint can be calculated by speed regulation), and / or at least one other operating variable VAF which can be a measurement or a setpoint.An operating variable of the electric machine is a quantity that characterizes the operation of the electric machine. This may include an electrical variable, such as the voltage, current, or power of the electric machine, a mechanical variable such as the position, speed, or acceleration of the rotor of the electric machine, etc. The operating variables may, for example, be the electrical rotation speed coe of the rotor (not shown) of the electric machine (alternatively, this variable may also be the norm of the magnetic flux in the phases), the voltage amplitude of the electric machine, the current amplitude of the electric machine, a temperature of the electric machine, etc. It is recalled that the electrical rotation speed coe of the rotor corresponds to the multiplication of the mechanical rotation speed com of the rotor by the number of pairs of poles of the synchronous electric machine MEL.In other words, we can write: coe=p.com with p the number of pairs of poles of the synchronous electric machine MEL. This electrical rotation speed makes it possible to determine the electrical frequency of the electric machine.

[0063] The COM control system comprises an OND inverter and a CAL computer. The COM control system can use, if necessary, a CAP sensor for the angular position and / or angular rotation speed of the rotor of the MEL electric machine. Such a sensor makes it possible to determine the angular rotation speed com of the electric machine either directly (in the case of a speed sensor) or by derivation (in the case of the position sensor). In addition, the COM control system can comprise means for measuring the currents (not shown) in the phases of the electric machine, for example current sensors.

[0064] The inverter OND is configured to route electrical energy between the DC source and the synchronous electrical machine MEL. More specifically, the inverter OND is configured to route electrical energy between the DC source and each phase of the stator of the synchronous electrical machine MEL. The DC source (for example a battery) provides a DC bus voltage, denoted VDc-

[0065] The OND inverter comprises a first input connected to the DC source, and three outputs, each connected to a corresponding phase of the stator of the synchronous electrical machine MEL. The OND inverter further comprises a second input electrically connected to an output of the CAL computer, such that the OND inverter is configured to route electrical energy between the DC source and the synchronous electrical machine MEL according to a switching control signal applied by the CAL computer to the second input of the OND inverter.

[0066] Preferably, the switching control signal may be such that the inverter OND routes electrical energy from the DC source to the synchronous electrical machine MEL so that the synchronous electrical machine MEL has an operation commonly called a "motor" and / or an operation commonly called a "generator". Conventionally, the inverter OND comprises several switching arms (not shown), preferably at least one switching arm for each phase of the electrical machine, to transform the DC signal from the DC source into an AC signal for the phases of the electrical machine MEL. Each switching arm comprises at least one controlled switch. Conventionally, each switch of the switching arms may be controlled by means of a pulse width modulation known by the acronym MLI (or PWM for Pulse Width Modulation). The inverter may be similar to that shown in [Fig.l].

[0067] The optional angular position or angular rotation speed sensor CAP can be configured to measure the angular mechanical position of a rotor (not shown) of the synchronous electrical machine relative to the stator (not shown), and to deliver an angular position signal (respectively angular rotation speed) representative of the measured value of the angular position (respectively angular rotation speed) of the rotor of the synchronous electrical machine.

[0068] Furthermore, the optional CAP angular position or angular rotation speed sensor can be configured to apply the angular position (respectively angular rotation speed) signal to a corresponding input of the CAL computer.

[0069] The method according to the invention is implemented from a voltage setpoint (also called reference voltage) of the phases of the electrical machine. This voltage setpoint can come from the CAL computer of the embodiment of [Fig. 14]. This is the expected voltage in each phase of the electrical machine, generally in a sinusoidal form.

[0070] The method according to the invention implements the following steps: - Detection of a saturation variation during a modulation period - Correction of the voltage setpoint - Determination of the switching setpoint - Control of the electric machine

[0071] The correction of the voltage setpoint is only applied in the case of a saturation variation during a modulation period. In other words, this correction is applied in an operating zone of the electrical machine in overmodulation and particularly in full-wave zones, but is not applied in the other operating zones, in particular the linear operating zone. The correction of the voltage setpoint makes it possible to reduce voltage errors in the overmodulation zone and in the full-wave zone.

[0072] These steps can be implemented in particular by a computer and a dedicated computer memory for real times. These steps are detailed in the rest of the description.

[0073] In the present application, the expression "modulation period" corresponds to the period of the PWM pulse width modulation method. In the figures, it is indicated TPWM. 1) Detection of a saturation variation

[0074] During this step, at least one modulation period is detected for each phase of the electrical machine for which the voltage setpoint is not entirely saturated at a single voltage by the voltage of the voltage source over said modulation period, and the instant of variation of the saturation of the voltage setpoint is detected. "not entirely saturated at a single voltage" implies that in the modulation period, there is at least one moment during which the voltage setpoint is saturated, and furthermore, the voltage setpoint is not restricted to a single saturation voltage. On the contrary, a modulation period saturated at (respectively 2 ~vdc ) over the entire duration of the modulation period is not detected. In other 2 terms, we detect at least one upcoming modulation period during which: - The voltage setpoint changes from an unsaturated state to a saturated state, - The voltage setpoint changes from a saturated state to an unsaturated state, or - The voltage setpoint changes from a first saturated state to a second state saturated.

[0075] In the first two situations, the instant of variation of the saturation corresponds to the instant of saturation.

[0076] Figures 3, 5 and 6 are examples of such modulation periods.

[0077] During this step, we do not detect the modulation periods fully saturated at a single saturation voltage (for example if the reference setpoint is saturated only at voltage + vdc). 2

[0078] The instant of variation of the saturation of the voltage setpoint corresponds to the instant within the modulation period, for which the voltage setpoint changes state, for example from a saturated state to a non-saturated state, or from a first saturated state to a second saturated state. This instant of variation of the saturation of the voltage setpoint is indicated by the reference IVs in FIGS. 3, 5 and 6.

[0079] According to an implementation of the invention, the detection implemented during this step can be carried out by comparing, for each future modulation period, the voltage setpoint with the saturation values ​​coming from the DC voltage source. 2) Correction of the voltage setpoint

[0080] During this step, for each modulation period detected in step 1 (and for each phase of the electrical machine), a correction of the voltage setpoint is applied, such that the corrected voltage setpoint is the weighted average of the voltage setpoint between two time intervals separated by said instant of variation of the saturation. In other words, for each modulation period detected in step 1, the voltage setpoint is corrected, by applying a new voltage setpoint value which represents the signal over the entire modulation period divided into two time intervals. The portion of the modulation period between the instant of variation of the saturation and the start or the end of the modulation period is called a time interval: each modulation period consists of two time intervals.The weighted average allows to take into account the contribution of each time interval (by discretizing each time interval separately), and consequently to make a voltage setpoint correction adapted to the contribution of each time interval, which allows to minimize the voltage error. The weighting of each contribution corresponds to the duration of each time interval. The idea is to modify the voltage setpoint Vref by taking into account the voltage error resulting from the conjunction of discretization and saturation: .

[0081] VrefCor=Vref-eVref

[0082] Where: Vrejcor is the corrected voltage setpoint, Vref the voltage setpoint (uncorrected), and £Vref is the error resulting from the conjunction of discretization and saturation.

[0083] Generally speaking, the corrected voltage setpoint can be written:

[0084] v _ v>-efn TpfrV„.jP2 Tn with Vrefçor the corrected voltage reference, TPWM ref Cor- TpwM the modulation period, VrefPi the voltage setpoint (saturated or not) during the first time interval of duration TVrefp2 the voltage setpoint (saturated or no) during the second time interval of duration Tn (with TPWM = TPl + TP2 )•

[0085] For the embodiment for which the voltage setpoint varies from an unsaturated voltage to a saturated voltage (overmodulation zone), the corrected voltage setpoint can be determined using the following formula: y _ v^fTm Tiv>sat with VrefCor the corrected voltage setpoint, le / Cor Tpwm T pwm the modulation period, VrefT„t the discretized saturated voltage setpoint during the saturated time interval Tsat of the modulation period, VrefT,wS,tt the discretized unsaturated voltage setpoint during the unsaturated time interval Tnosat of the modulation period (with TPWM — Tsat + TnoSat).

[0086] Advantageously, the unsaturated voltage setpoint may be the average of the voltage setpoint during the unsaturated time interval of the modulation period.

[0087] Figures 7 and 8 illustrate, schematically and in a non-limiting manner, this correction for this embodiment. Figures 7 and 8 are similar to Figures 2 to 6, and correspond to the configuration of Figure 3. For the modulation period T PWM, a variation in saturation is detected, at time Ivs. The modulation period is then divided into two time intervals: an unsaturated time interval of duration TnoSat and a saturated time interval Tsat. The unsaturated time interval is discretized to deduce a discretized voltage setpoint Vrefr, and the saturated time interval is discretized to deduce a discretized voltage setpoint VrefPwi which corresponds to the saturation voltage.Then, the corrected setpoint is determined by a weighted average of these discretized voltage setpoints VrefCor- Thanks to this weighted average, the voltage error becomes zero because the surfaces above and below the curve are identical.

[0088] For the embodiment for which the voltage setpoint varies from a first saturated voltage to a second saturated voltage (full wave zone), the corrected voltage setpoint can be determined using the following formula: tr _ vrefr^ Twith VrefCor the corrected voltage setpoint, TPWM r refCor “ TJ J 1 PWM the modulation period, VrefTmû the saturated voltage setpoint during the first saturation time interval of the modulation period, V ref T the saturated voltage setpoint during the second saturation time interval TSat2 of the modulation period (with T pwm — Tsati + ^.%r2), the first saturation time interval being separated from the second saturation time interval at the switching variation instant.

[0089] Figures 9 and 10 illustrate, schematically and in a non-limiting manner, this correction for this embodiment. Figures 9 and 10 are similar to Figures 2 to 6, and correspond to the configuration of Figure 5. For the modulation period T pwm, a variation in saturation is detected at time IVs. The modulation period is then divided into two time intervals: a first saturated time interval of duration TSati and a second saturated time interval Tsat2. The first saturated time interval is discretized to deduce therefrom a discretized voltage setpoint VrefT^l} 9ui which corresponds to the saturation voltage -vdc, and the second saturated time interval is discretized to deduce therefrom a discretized voltage setpoint VrefTSilO which corresponds to the saturation voltage . Then, the setpoint is determined 2 corrected by a weighted average of these discretized voltage instructions Vrefcor. Thanks to this weighted average, the voltage error becomes zero because the surfaces above and below the curve are identical. 3) Determination of the switching setpoint

[0090] During this step, a switching signal is determined for each phase of the electrical machine for each switching arm of the inverter as a function of the corrected voltage setpoint determined in step 2. In other words, the switching duration of the switching arms of the inverter is determined in such a way (also called duty cycle) as to obtain the corrected voltage setpoint.

[0091] Advantageously, the switching signal can be determined in a conventional manner, using the PWM method.

[0092] In the full-wave case, the ideal control voltage setpoint, it is recalled, is a square wave of value _ vdc during one time interval of the electrical period and of value . vdc during the other time interval. However, in ap-+ 2 By applying the corrected voltage setpoint, an additional unwanted pulse is introduced. This pulse is the direct consequence of the correction: we move from applied commands which are only worth the minimum or the maximum (therefore commands which impose either _ vx , or , vdc during the entire period 2 + 2 PWM) to an applied command calculated by an average: it is therefore an intermediate value. This corrected command avoids generating control voltage errors but introduces two additional switchings (changes from min to max or vice versa). However, these switchings generate additional losses in the inverter.

[0093] Figures 11 and 12 illustrate, schematically and in a non-limiting manner, the switching setpoint Vcom of phase a of the electrical machine as a function of the electrical angle Se in °, in the full-wave case, with a correction of the voltage setpoint in the modulation period ranging from 34° to 51° of the electrical angle. [Fig.12] is an enlargement of [Fig.1 1]. These figures show the additional induced pulses.

[0094] According to an embodiment of the invention, making it possible to limit the number of pulses and therefore the losses in the inverter, a switching setpoint can be determined by implementing the switching interval at one end of the modulation period. The switching interval is the duration during which the switching is implemented. This interval is noted Tcom in [Fig. 12]. Conventionally and as illustrated in Figures 11 and 12, the switching interval is implemented at the center of the modulation period. This embodiment proposes to shift the switching interval to the beginning or the end of the modulation period. Thus, continuity of the state of the switch is allowed, which limits the number of switchings and the losses in the inverter.

[0095] According to a first example, the switching interval can be started at the beginning of the modulation period when the voltage setpoint varies between maximum saturation and minimum saturation. Thus, the switching setpoint remains at the maximum saturation level at the beginning of the modulation period.

[0096] According to a second example, the switching interval can be terminated at the end of the modulation period when the voltage setpoint varies between minimum saturation and maximum saturation. Thus the switching setpoint is always at the level of maximum saturation at the end of the modulation period.

[0097] [Fig. 13] illustrates, schematically and in a non-limiting manner, the correction of the switching setpoint for the example of [Fig. 12]. In this case, the switching interval Tcom is shifted to the end of the modulation period. The corrected switching setpoint obtained is noted Vcom>cor, and makes it possible to limit the number of switching operations. 4) Inverter control

[0098] During this step, each switching arm of the inverter is controlled by applying for each phase of the electrical machine the switching setpoint determined in step 3. In other words, an opening and / or closing setpoint is applied to each switch of the switching arms of the inverter so as to correspond to the switching setpoint. In this way, the inverter generates a machine control current signal close to a sinusoidal signal.

[0099] Furthermore, the invention relates to a system for controlling a synchronous electrical machine comprising an inverter provided with switching arms, a computer and a memory configured to implement the steps of the control method according to any one of the preceding variants or according to any one of the com combinations of previous variants, to apply the switching setpoint to the inverter switches.

[0100] Preferably, the electrical machine may be a permanent magnet assisted synchronous-reluctant machine or a permanent magnet synchronous machine.

[0101] Advantageously, the switching arms of the inverter may comprise wideband semiconductors of the WBG ("wide band gap") type power modules, which generate significantly fewer losses at each switching period. Indeed, the present invention allows a suitable switching frequency for these switches. Alternatively, the power modules may be of the IGBT or MOSFET type.

[0102] As goes without saying, the invention is not limited to the embodiments described above as examples; on the contrary, it encompasses all variant embodiments.

Claims

Claims

1. Method for controlling an electrical machine (MEL) by means of pulse width modulation from a voltage setpoint of the phases of said electrical machine, said electrical machine being controlled by an inverter (OND) provided with a plurality of switching arms (BC) and powered by a direct voltage source (DC), characterized in that the following steps are implemented: a. At least one modulation period (TPWM) is detected for which said voltage setpoint is not entirely saturated at a single voltage by the voltage of the direct voltage source (DC) over said modulation period (TPWM), and the instant of variation of the saturation (Ivs) of said voltage setpoint is detected; b.For each detected modulation period (T PWM), a correction of said voltage setpoint is applied, such that the corrected voltage setpoint is the weighted average of the voltage setpoint of two time intervals separated by said instant of variation of saturation (Ivs) over the modulation period; c. A switching signal is determined for each switching arm (BC) of said inverter (OND) as a function of said corrected voltage setpoint; and d. Each switching arm (BC) of said inverter (OND) is controlled by applying said determined switching setpoint.

2. Control method according to claim 1, in which, for each modulation period (TPWM) detected, and when said voltage setpoint changes from a first saturated state to a second saturated state or when the voltage setpoint is overmodulated, a switching setpoint is determined, by carrying out the switching interval at one end of the modulation period.

3. Control method according to claim 2, in which the switching interval is started at the beginning of the modulation period when the voltage setpoint varies between maximum saturation and minimum saturation.

4. Control method according to claim 2, in which the switching interval is terminated at the end of the modulation period when the voltage setpoint varies between minimum saturation and maximum saturation.

5. Control method according to one of the preceding claims, wherein when the voltage setpoint varies from an unsaturated voltage to a saturated voltage, said corrected voltage setpoint is determined by means of the following formula: V — with VrefCor the corrected voltage setpoint refCor TPWM, TPWM the modulation period, VrefTmt the saturated voltage setpoint during the saturated time interval Tsat of the modulation period, VrefT,mSr,t the unsaturated voltage setpoint during the unsaturated time interval TmSat of the modulation period, the saturated time interval being separated from the unsaturated time interval at said switching variation instant.

6. A control method according to claim 5, wherein the unsaturated voltage setpoint is the average of the voltage setpoint during the unsaturated time interval of the modulation period.

7. Control method according to one of claims 1 to 4, wherein when the voltage setpoint varies from a first saturated voltage to a second saturated voltage, said corrected voltage setpoint is determined by means of the following formula: V — V,'etT«n with VrefCor the corrected voltage setpoint refCor TPWM, TPWM the modulation period, Vrefr the saturated voltage setpoint during the first saturation time interval Tsat} of the modulation period, V refT^ the saturated voltage setpoint during the second saturation time interval TSan of the modulation period, the first saturation time interval being separated from the second saturation time interval at said switching variation instant.

8. Control system for an electrical machine (MEL) comprising an inverter (OND) provided with switching arms, a computer (CAL) and a memory configured to implement the steps of the control method according to one of the preceding claims.

9. A control system according to claim 8, wherein said electric machine (MEL) is a permanent magnet-assisted synchronous-reluctant machine or a permanent magnet synchronous machine.

Images