Control of an electrical machine with harmonic suppression in several quantities by counter signal generation in space vector representation

Abstract

Control device for controlling an electric drive, which has an inverter (INV) and a multi-phase electric machine (EM) connected thereto, wherein the control device has a basic control loop (GRK) with a deviation detection (RA), an actuator (IRT) and with a controlled system (RS) which comprises the inverter (INV) and the electric machine (EM), wherein in the feedback path (FB_G) of the basic control loop (GRK), the current intensity of the multi-phase current (I_U, I_V, I_W) flowing between the inverter (INV) and the electric machine (EM) is fed back as an actual value (I_EM) for deviation detection (RA), wherein the basic control loop (GRK) refers to a fundamental angular frequency (ω_G) and wherein the control device has a harmonic detection (OE, RT_O) which is connected downstream of the feedback path (FB_G) of the basic control loop (GRK) and is designed to determine the harmonic component (I_O) in a power supply signal (U_P),which supplies the inverter (INV), and has a counter-signal generator (GS_RZ) which is connected downstream of the harmonic detection (OE) and is configured to generate a counter-signal (GS) which is fed into the control variable (RZG) of the actuator (IRT) via a feed (EGS) connected upstream of the actuator (IRT), wherein the counter-signal generator (GS_RZ) is configured to generate the counter-signal (GS) provided in a space vector representation from the harmonic component (I_O) by adjusting the two vectors of the space vector representation both to minimize the harmonic component (I_O) in the power supply signal (U_HV) and to reduce a harmonic component in the torque of the electric machine (EM), wherein the counter-signal (GS) provided in a space vector representation generates two mutually perpendicular voltage vectors,wherein a first of these vectors (U_q(O)) represents at least a portion of the magnetic flux density of the harmonic and a second of these vectors (U_d(O)) indicates at least a portion of the torque of the harmonic, wherein the first and the second vectors are adjusted according to a minimization of the harmonic component in the torque according to a minimization of a reactive component, wherein the reactive component is provided in the harmonic component (I_O) in the actual value (I_EM), or the reactive component corresponds to the harmonic component (I_O) in the actual value (I_EM).

Description

[0001] It is known, for example, to equip electric vehicles with an electric drive, with an electric motor generating the traction power. This not only generates a rotating field with the desired speed (or, in the case of asynchronous machines, a rotating field for the desired rotor speed), but also results in harmonic components due to numerous effects. These harmonic components are undesirable and generate losses, noise, and similar side effects.

[0002] To suppress the harmonic components, a counter signal can be generated that is inverted to the interfering harmonic, so that when this counter signal is injected, at least partial cancellation occurs. This active cancellation is based on a counter signal with the correct amplitude and, above all, the correct phase (relative to the interfering signal) to cancel it out. By suppressing harmonics in the electrical control signal or in the DC supply current, a DC link capacitor can be relieved and excessive voltage values ​​can be avoided.

[0003] The publication “Selective current harmonic suppression for high-speed PMSM based on high-precision harmonic detection method”, LIU, Gang [et al.], IEEE Transactions on Industrial Informatics, Vol. 15, 2019, No. 6, pp. 3457-3468 focuses on the above-mentioned problem of harmonics in phase currents of an electric drive and proposes a control loop for harmonic suppression that uses several synchronous rotary systems to represent the harmonics in the phase currents.

[0004] Harmonic suppression can be used for the optimized design of electrical and electronic components. The subsequently published document DE 10 2022 210 516 A1 describes a space vector-based method for suppressing harmonics in electrical drive signals by injecting a corresponding counter signal.

[0005] In addition, harmonics arise not only in the control or supply voltage, but also in the torque curve, which are particularly detrimental in terms of noise generation. Regarding the problem of harmonics in the torque curve, the publication US 2018 / 0127022 A1 describes an active steering device in which torque ripple is reduced using a compensation signal generated using a lookup table.

[0006] It is therefore an object of the invention to show a possibility with which electrical and mechanical harmonics can be actively suppressed in a simple manner.

[0007] This object is achieved by the subject matter of claim 1. Further properties, features, embodiments and advantages result from the dependent claims, the description and the figure.

[0008] It is proposed to generate a counter signal in space vector representation in a basic control loop of an electrical machine. The space vector representation is based on two vectors that represent different quantities (such as magnetic flux density and torque), whereby these quantities have a phase shift to each other. The vectors can be viewed as two quantities between which there is a physical relationship (in particular based on electromagnetic induction), which provides a time derivative that is reflected in the phase shift. A first of the vectors (such as a d-component [direct component] or α-component) represents a magnetic flux density, while the second vector (such as a q-component [quadrature component] or β-component) represents the torque in the electrical machine (if necessary, neglecting cross-dependencies due to reluctance effects). The second vector, for example, indicates a direct effective quantity (i.e.an in-phase quantity, such as torque or a torque-generating current or voltage component), while the first phasor represents a phase-shifted reactive component (magnetic flux density). This applies to a simplified model that does not take reluctance effects into account. In models that take reluctance into account, the two phasors do not exclusively represent the respective physical quantity (flux density or torque). In representations that also take reluctance effects into account, one phasor represents a portion of the flux density and a portion of the torque, while the other phasor represents a different portion of the flux density and a different portion of the torque.Since the two pointers represent different parts of the flux density and different parts of the torque, the pointers are orthogonal with respect to the two quantities, so that both quantities can be adjusted with both pointers (although interlaced or not completely separate from each other).

[0009] The second pointer refers (in a simplified model without taking reluctance into account) to the torque of the harmonic in the counter signal, i.e. oscillating forces caused by the harmonic. The first pointer relates (in a simplified model without taking reluctance into account) to an electrical quantity, such as a harmonic current component in the phase current of the electrical machine or a harmonic voltage component or harmonic current component in the DC supply voltage of the inverter. Taking reluctance into account, both pointers represent the torque and the electrical quantity, but in different proportions, so that with the two pointers there are two setting options (degrees of freedom) for setting the two quantities (harmonic torque and harmonic component of the electrical quantity).Minimizing the electrical harmonics (minimizing the phase current or the harmonic component in the supply voltage or current, generally the harmonic component in a power supply signal [DC] of the inverter or in an intermediate circuit) can therefore be a first objective when generating the countersignal, while minimizing the (oscillating) harmonic component in the torque as a mechanical harmonic component can be a second objective. A first control objective can therefore be minimizing the ripple load of a DC link supplying the inverter, while a second control objective can be minimizing the ripple in the torque. Both objectives can be pursued simultaneously due to the phasor representation (comprising two phasors) of the countersignal.The control objectives are, in particular, the control objectives of a harmonic control loop that includes the counter-signal generation (as the actuator). Minimization does not refer to global, maximum optimization, but also to partial minimization or reduction, not necessarily to zero.

[0010] It was thus recognized that when generating a countersignal in vector representation, both vectors can be used to optimize (minimize) the two quantities (i.e., an electrical harmonic quantity and a mechanical harmonic quantity), so that both quantities can be reduced (at least partially) separately from each other during the generation of the countersignal, since both vectors can be adjusted separately (for reduction in both target directions). As an additional constraint when generating the countersignal, the amplitude of the countersignal can be minimized (or at least reduced).

[0011] A control device for controlling an electric drive, in particular a traction drive of a vehicle, is described. The drive has an inverter and a multi-phase electric machine connected to it. The electric machine has a stator whose windings are connected to the inverter. The inverter is designed to generate a rotating magnetic field in the stator by appropriately controlling the windings. The inverter has a DC voltage side that is connected to a DC supply connection or to a DC supply voltage. Harmonics are generated in the DC supply voltage, among other things, by the switching events of the inverter. The inverter generates a multi-phase current ("phase current") that causes the rotating field in the stator. Harmonics also arise in the current due to the operation of the electric machine, the switching events of the inverter, and possiblydue to other effects, harmonics. These current harmonics or the harmonics in the DC supply voltage are reduced by a counter signal. However, the counter signal is preferably not injected directly into the supply voltage or the phase current, but is combined with signals present in the control device to reduce the respective harmonics. Reduced harmonics relieves the DC link (or the DC supply voltage) in particular with regard to alternating loads. This leads in particular to a DC link capacitor that is less loaded (with alternating components). Furthermore, oscillating effects in the torque caused by harmonics are also reduced. This reduces noise generation.

[0012] The control device is preferably designed for vector control.

[0013] The control device has a basic control loop with deviation detection, an actuator, and a controlled system. The controlled system comprises the inverter and the electric machine. The actuator preferably has an inverter control. This generates the switching signals for the switching elements of the inverter. The switching signals are preferably pulse-width modulated signals. The input variable of the actuator is a controlled variable, which is preferably present in space vector representation. The controlled variable is in particular a voltage signal with a d and a q component or outputs a voltage signal with an active component (torque - q component), corresponding to the second vector of a space vector transformation, and with a reactive or quadrature component (magnetic flux density or excitation - d component), corresponding to the first vector of a space vector transformation.The phasors do not necessarily have to represent pure active and pure quadrature quantities; rather, a different weighting of these components, i.e., a phase offset > 0 between the phasors of the space phasor representation, is sufficient to represent the manipulated variable (complex, i.e., with phase information). In simplified models without considering reluctance effects, one of the phasors is directly assigned only to the active component, and the other phasor is directly assigned only to the reactive component. Taking reluctance into account, one of the phasors is directly assigned to both the active component and the reactive component, and the other phasor is directly assigned to both the active component and the reactive component.However, since the assignment of one pointer to the two components differs from the assignment of the other pointer to the two components, both components can be adjusted using the pointers (taking into account the cross-dependence of the pointers to the components). By adjusting the two pointers, the two components can be adjusted and, in particular, minimized, with or without taking cross-dependence and reluctance effects into account.

[0014] The actuator is preferably designed for inverse space vector transformation or has a corresponding transformation device. The actuator generates the switching signals in the time domain (particularly pulse-width modulated). These switching signals are suitable for controlling the (multiphase) inverter. The actuator is connected to the inverter for control purposes. The switching signals relate to the various phases of the inverter (U, V, W). The pulse-width modulation of the switching signals is carried out according to the control variable, which is represented in space vector form.

[0015] The basic control loop (like every control loop) has a feedback path. The feedback path feeds back the current (as the controlled variable) that flows between the inverter and the electrical machine, or the current that flows in the electrical machine or is output by the inverter. The current is fed back as a signal or variable that reflects this flowing current. The current (or the variable that reflects it) is fed back as an actual value (or as a controlled variable). A measuring device can be provided that records the actual value (of the controlled system, inverter, or electrical machine), in particular the aforementioned current. The current is fed back for the purpose of deviation detection, i.e., for deviation detection. Deviation detection is set up to determine the error (or difference) between the fed-back actual value and a setpoint. The result is output as a control difference, for example to a control element of the basic control loop.If the setpoint is in space vector representation, a space vector transformation is provided in the feedback path, which transforms the (multi-phase) actual value in time domain representation into a space vector representation of the actual value. If the setpoint is entered in time domain representation at an input (setpoint input), a further space vector transformation for the setpoint can be provided between the deviation detection and the input. If the setpoint is in time domain representation, the actual value in time domain representation can also be passed on to the deviation detection. A subtraction element can be provided in the feedback path to subtract a counter signal so that a counter signal fed into or modulated onto the controlled variable can take effect and is not undesirably suppressed by the control device (e.g. by a control element).Furthermore, not only the setpoint (speed, torque) can be entered at the setpoint input, but also operating parameters or secondary condition parameters can be entered at this input, for example the current temperature of the electric machine and / or the inverter and / or the level of the DC voltage that forms the supply voltage of the inverter, or even a ripple strength that the DC voltage has. The input can therefore have a setpoint input for speed and torque as well as an actual value input for (actual) operating parameters. However, only the setpoint (speed, torque) of the input is entered into the deviation detection; the operating parameters (actual variables of secondary conditions) are passed on, in particular, to a counter-signal generator of the control device.

[0016] The basic control loop is referenced to a fundamental angular frequency. This applies in particular to the basic control loop variables that may be provided in space vector representation (setpoint, actual value, error or deviation or control difference, output variable before the control elements of the basic control loop, etc.). The relevant variables are referenced to a rotating phasor representation, with the angular velocity of this rotation corresponding to a fundamental wave. The space vector representation can have a phasor whose angular frequency corresponds to the fundamental angular frequency. The angular velocity depends in particular on the electrical angular velocity of the electrical machine (or the power signals output by the inverter) and thus also on the rotational speed of the electrical machine.

[0017] The control device has a harmonic detection system. This is configured to detect a harmonic component from the actual value (present in the time domain or as a space vector representation). The harmonic detection system is configured to detect at least one harmonic component from the actual value by Fourier analysis (in particular discrete Fourier analysis), by bandpass filtering, or by integrating a signal resulting from modulation with at least one harmonic frequency. The harmonic component refers to harmonics relative to the fundamental wave and, in particular, comprises at least one alternating component with a frequency that corresponds to n times the fundamental wave frequency, where n = an integer and > 1.The harmonic detection can be set up to generate the harmonic component by transforming the actual value (multi-phase) into a space vector representation, wherein the space vector representation rotates at a harmonic angular frequency (relative to the fundamental angular frequency of the basic control loop. From this space vector representation, which refers to a harmonic angular frequency (“harmonic space vector representation”), the strength of the harmonic that has the harmonic angular frequency can be determined by absolute value formation. The phase position is determined based on the phase position of the vector of the space vector representation. To determine harmonics with different harmonic angular frequencies, it can be provided that the actual value is transformed into several space vector representations that have the different harmonic angular frequencies. The harmonic detection is preferably set up for this purpose.The harmonic detection device can be connected downstream of a measuring device (phase current measuring device) that detects or determines the actual value (also for the feedback path). The harmonic detection device is connected downstream of the feedback path of the basic control loop in such a way that the harmonic detection device receives the actual value transmitted via the feedback path. The harmonic detection device receives the actual value (phase current of the inverter / electrical machine) as a signal in the time domain (multiphase) or in space vector representation.

[0018] The control device has a counter-signal generator configured to generate a counter-signal. The counter-signal is complementary to the harmonics and can thus at least partially compensate for them (when they are combined). The counter-signal is, in particular, 180° out of phase with the detected harmonics or the signal that represents the harmonics (including phase information). The counter-signal generator receives the harmonics or a signal that represents them (preferably including phase information). For this purpose, the counter-signal generator is connected downstream of the harmonic detection. Within the control device, the counter-signal is combined with the harmonics (or with a signal containing them). The counter-signal is combined with the control variable of the actuator. This is carried out by means of a feed. The control variable preferably corresponds to a voltage signal that is input to the actuator.The voltage signal is preferably in space vector representation, but can also be in the time domain. The counter signal is preferably in space vector representation if the controlled variable is in space vector representation. The counter signal is preferably in the time domain if the controlled variable is in the time domain. For this purpose, the representation of the counter signal or the controlled variable can be adapted by a transformation between space vector representation and representation in the time domain. A control element is preferably connected downstream of the deviation detection, which in turn is followed by the feed-in. The feed-in is preferably connected upstream of the actuator. The feed-in combines the counter signal with the controlled variable and thus at least partially compensates for harmonics present in the controlled variable.

[0019] The counter-signal generation is configured to generate the counter-signal from the harmonic component determined by the harmonic detection. The counter-signal is provided in a space vector representation, either by the harmonic detection determining the harmonic component in a space vector representation, or by converting the signal emitted by the harmonic detection (which indicates the harmonic component) into a space vector representation. The space vector representation of the harmonic component is based on vectors that rotate at a harmonic angular frequency, i.e., at an angular frequency greater than the fundamental angular frequency.

[0020] The counter signal is represented in a space vector representation based on two rotating vectors. The space vector representation, in particular, provides two rotating axes with a fixed angle to each other, which are perpendicular (or generally angularly offset). In the case of mutually perpendicular axes, the space vector representation refers to a complex representation, where the two axes correspond to the real and imaginary components of the complex representation.

[0021] The pointers are phase-shifted by a fixed angle, in particular by 90°. This allows for several degrees of freedom in generating the counter signal, particularly since the magnitude and phase of the counter signal can be represented by a virtually arbitrary linear combination of these pointers. This is also the case when, taking reluctance effects into account, both pointers refer to an active and a reactive component (of the harmonic counter signal), but the pointers are assigned to the active component and the reactive component in different ways. Even with such a cross-dependence of the pointers and the components, the two components can be adjusted and, in particular, minimized.

[0022] Multiple optimization objectives can be achieved when generating the counter signal. One optimization objective can be to minimize the harmonic component, i.e., the harmonic component delivered by the feed-in (as a control variable to the actuator). Another optimization objective can be to minimize the harmonic component in the torque of the electric machine (or a harmonic component in the DC supply voltage of the inverter). The counter signal generation is configured for optimization according to these two optimization objectives. The counter signal generation can be configured to execute a corresponding optimization algorithm. Alternatively, pointer parameters optimized according to these optimization objectives are stored in the counter signal generation for various harmonic amplitudes, harmonic phases, and / or harmonic angular frequencies ("optimization input values").The phasor parameters, in particular, represent at least one phase angle and at least one amplitude for the two phasors of the countersignal. The countersignal generation is preferably configured to retrieve the associated optimized phasors or phasor parameters for the optimization input values. The optimization input values ​​and the associated phasor parameters can be stored in a lookup table in the countersignal generation or can be stored as parameters of an approximate equation, with the countersignal generation preferably configured to calculate the optimized phasors (as an approximation) using the approximate equation. Both variants require less computing power than real-time optimization.

[0023] In addition to the deviation in the basic control loop, the counter signal is also crucial for the control variable. Thus, by modifying the control device for harmonics (harmonic detection, counter signal generation), harmonic suppression can be achieved not only in the control according to the basic control loop but also in the torque of the electric machine. In addition to the harmonic component in the actual value, the counter signal generation receives the harmonic components in the torque of the electric machine, which result in particular from torque detection or from deriving the harmonic component in the torque from an operating variable of the inverter and / or the electric machine.

[0024] A space vector representation based on the fundamental angular frequency can be related to the rotation of the rotor and be static relative to the rotor. However, the counter signal and the harmonic component rotate at a multiple relative to the rotor, particularly when the counter signal is fed in. If the counter signal or the harmonic component is n times the fundamental frequency, then the counter signal and the harmonic component are present as a harmonic angular frequency, which corresponds to n times the fundamental angular frequency. The counter signal, particularly if present in space vector representation, can be modulated by the desired n-fold, i.e., with a cosine or sine signal whose frequency corresponds to the harmonic angular frequency. A modulator (for signals in the time domain or for space vector representations) can be provided for this purpose.When the counter signal is fed in as a time-domain signal, the counter signal is combined as a signal modulated with the harmonic angular frequency. This means that the harmonic component to be reduced and the counter signal have the same frequency, namely the frequency corresponding to the harmonic angular frequency, so that with appropriate phasing, at least partial cancellation of the harmonic component can occur.

[0025] When calculating the harmonic component in space vector representation, it must be considered that this representation refers either to the rotor or to the stator. Differences arise if the fundamental space vector representation of the respective manipulated variable is stator-fixed or already rotates with the rotor and is thus rotor-fixed, depending on the type of space vector representation (d,q representation or α,β representation). Here, the harmonic component is captured in space vector representation by mapping the actual value (represented in a space vector representation based on the fundamental angular frequency) using a space vector representation based on the harmonic angular frequency.

[0026] The harmonic detection device can also be designed as a high-pass filter or band-pass filter. This filter is configured to block the fundamental angular frequency and allow the harmonic angular frequency to pass. The filter is particularly designed to attenuate the fundamental angular frequency by at least five, ten, 20, or 30 decibels more than the harmonic angular frequency, or frequencies that are more than a factor of 2 or its inverse, or more than a factor of 4 or its inverse, of the harmonic angular frequency. The filter preferably displays the relevant phase position in addition to the signal strength (amplitude) of the harmonic(s). Furthermore, the harmonic detection device can also be designed as a spectral analysis device that implements, for example, a Fourier transformation.The spectral analysis device is designed to extract components of the harmonic angular frequency, in particular together with the phase position of these components. The harmonic detection and counter-signal generation are designed to process one or more different harmonics.

[0027] The harmonic component is mapped to the counter signal taking into account an angular relationship between the control variable (d,q components of the voltage) in the controlled system and the controlled variable (ie actual value or multi-phase phase current).

[0028] The space vector representation of the counter signal is thus preferably represented by two rotating axes that are fixed in angle to one another, are not parallel and in particular form an angle of 90°. The space vector representation of the counter signal can thus be particularly complex, with a real axis and an imaginary axis offset by 90°. The counter signal calculation provides a first reduction condition, namely the minimization of the harmonic component in a power signal, which can be fulfilled with only one of the axes of the space vector representation or with both axes, but while maintaining a (further) degree of freedom. The "power signal" refers to the actual value or the (multi-phase) phase current of the electrical machine or inverter, or the power supply signal of the inverter, which represents the DC voltage supply or DC current supply of the inverter. The power signal is an electrical quantity.This results in an underdetermined system. The two axes provide two degrees of freedom, so that the first degree of freedom can be used to minimize the harmonic content in the power signal as a first optimization objective, while the other axis can be chosen essentially freely, for example according to a further optimization objective. The further optimization objective can be to minimize the harmonic content present in the torque of the electrical machine. The further optimization objective therefore relates to a mechanical quantity, for which an electrical equivalent can, however, exist, on the basis of which the optimization is carried out. Both optimization objectives can each affect both phasors of the space vector representation, or one optimization objective can affect one phasor while the other optimization objective affects the other of the two phasors in the space vector representation.

[0029] It can be provided that the counter-signal calculation is configured to map a value of one of the phasors of the counter-signal calculation according to the first reduction condition or according to the first optimization objective, while a value source specifies the associated value of the other axes. This is preferably the case when one phasor can be assigned to only one mechanical quantity (torque) and the other phasor can be assigned to only the electrical quantity (current, voltage). In particular, when taking into account cross-dependencies between the two phasors on the one hand and the two quantities (mechanical, electrical) on the other hand, it can be provided that a phasor value pair having a first phasor value and a second phasor value is assigned to different harmonic value pairs, each of which has a harmonic value of the power signal and a harmonic value of the torque.The phasor values ​​indicate the values ​​of the phasors for which the harmonic content is minimized or reduced when the corresponding harmonic value pair is present. This assignment can be implemented using a multidimensional lookup table, an approximation function, or similar. The assignment can be based on calculations, in particular model-based calculations related to the inverter and / or electrical machine used, or can be the result of tests based on the electrical machine / inverter in question, or a combination of these, such as test results interpolated by model-based calculations.

[0030] In addition to the basic control loop, the control device can thus have a harmonic control loop that generates the counter signal according to two control objectives. This counter signal is represented in space vector form (with two vectors). The actual parameter of the harmonic control loop is the harmonic component (or its strength of the (electrical) power signal) or the harmonic component derived from it in the torque of the electric machine. The actual parameter is thus a pair of values ​​or a vector with two entries. The counter signal generation serves as an actuator within the harmonic control loop, whereby, for example, any harmonic component deviating from 0 is considered an error, corresponding to a deviation detection for the harmonic. A harmonic space vector transformation of the actual value (if present in the time domain) can be considered a deviation detection (deviation from 0) for the harmonic.The input of the counter signal into the fundamental control loop (or its control variable) serves as the (executive part of) the controller(s) of the harmonic control loop. Further features, particularly related to the space vector representation of the harmonic and fundamental oscillations, are described in the further description and in . Fig. 1 shown.

[0031] The feedback path of the basic control loop preferably has a subtraction element. This is configured to subtract the harmonic component from the actual value transmitted by the feedback path, with the resulting signal (i.e., the difference) being passed on by the subtraction element within the feedback path. In this way, the influence of the (detected) harmonic is suppressed within the feedback path, so that the basic control loop does not change the manipulated variable due to the harmonic components. Instead of subtracting the harmonic component from the controlled variable, the subtraction element can also be configured to subtract the countersignal from the controlled variable.

[0032] The subtraction element operates in a space vector representation; therefore, subtraction is performed with quantities that are present in space vector representation. The actual value (or the relevant feedback quantity) is input to the subtraction element in a space vector representation. The harmonic component or the countersignal is also fed to the subtraction element in a space vector representation. The harmonic component, for example in the actual value (or in the power signal), can be recorded in space vector representation, or can be recorded as a time signal and transformed into space vector representation by a harmonic-space vector transformation (before being fed to the subtraction element). The subtraction element receives the actual value (generally: the feedback quantity) in a space vector representation based on rotation at the fundamental angular frequency.Since the space vector representation of the harmonic component or the counter signal for subtraction from the controlled variable should also refer to the fundamental angular frequency, the harmonic component or the counter signal is passed to the subtraction element modulated with the harmonic angular frequency. A modulator can therefore be connected upstream of the subtraction element, which modulates the feedback variable with the harmonic angular frequency. The subtraction element thus receives the harmonic component or the counter signal in space vector representation, based on the fundamental angular frequency, but modulated with the harmonic angular frequency, in order to account for the fact that the feedback path passes on the controlled variable as a space vector representation based on the fundamental angular frequency, and that the harmonic component has a comparatively higher frequency, namely the harmonic angular frequency.If the counter signal is fed in instead of the harmonic component, then preferably a version of it rotated by 180 degrees is fed to the subtraction element instead of the fundamental source signal, or a version of it multiplied by minus 1.

[0033] The subtraction element is located in a section of the feedback path of the basic control loop, where signals are transmitted in space vector representation. The space vector representation refers to the fundamental angular frequency. The harmonic component is also subtracted in space vector representation in the subtraction element, with this space vector representation based on the harmonic angular frequency.

[0034] It can be provided that the harmonic space vector transformation transforms the actual value (or power supply signal) present in the time domain into a space vector representation in order to forward it, in particular to the subtraction element or the counter signal generator. Since the output of the harmonic space vector transformation is ultimately fed into the basic control loop, whose space vector transformation also refers to the fundamental angular frequency, the harmonic component in the power signal is transformed according to the ratio of the harmonic angular frequency to the fundamental angular frequency.The latter serves, on the one hand, to subtract space vector representations whose coordinate system rotates at the same angular frequency (the fundamental angular frequency), while simultaneously taking into account that the harmonic component rotates faster in space vector representation than the signal in the fundamental control loop, namely at the harmonic angular frequency. The latter is achieved by the space vector transformation, which is based on a ratio of the harmonic angular frequency to the fundamental angular frequency (both for the positive harmonic angular frequency and the negative harmonic angular frequency, which applies to a counter-rotating harmonic system).

[0035] Further embodiments provide for a two-part linear controller to be provided between the deviation detection and the subsequent input of the counter signal. The controller has, in particular, a first and a second controller part. The first controller part relates to a first axis of the space vector representation, and the second controller part relates to a second axis of the space vector representation. In particular, the first controller part can relate to a real axis, and the second controller part to an imaginary axis, of the space vector representation of the signal that is passed on in the controlled system. The controller parts therefore also work with signals that are represented in space vector representation. This simplifies the control process. The controller parts are preferably each linear controllers, such as PID controllers. The controller parts are preferably not interleaved and control individually. The controller parts are, in particular, each step controller.

[0036] A limiter can be connected downstream of the counter-signal generation or calculation. This limiter can reduce the amplitudes of the space vector representation of the counter-signal. The limiter is preferably connected upstream of the modulator, via which the limited signal is fed to the counter-signal input. Alternatively, the limiter can be connected downstream of the modulator to provide amplitude limitation for the modulated counter-signal (in space vector representation). The limiter also operates with signals provided in space vector representation, particularly in a space vector representation that refers to rotation at the fundamental angular frequency.

[0037] A sensor, in particular an angle sensor, can be arranged on the electrical machine or, alternatively, a unit that determines the current rotor angle of the electrical machine, for example from the power signals (current and / or voltage signals) with which the electrical machine is operated. The angle signal emitted by the sensor can be transmitted to an input of the harmonic detection. The harmonic detection is configured to relate the power signal of the electrical machine or the power signal emitted by the inverter to the angle signal in order to output a space vector representation of the harmonic component together with the correct angle. Preferably, the angle signal is adapted to the variables with which the harmonic detection operates, wherein the harmonic detection receives a modulated signal.Therefore, at least one harmonic modulator can be provided to modulate the angle signal in order to adapt it to the rotation of the space vector representation of the harmonic detection. In particular, two harmonic modulators can be provided to modulate the angle signal, on the one hand, with a frequency corresponding to the harmonic angular velocity plus 1 (corresponding to the rotation of the basic system or the rotor) or plus 2π (corresponding to the rotation of the basic system or the rotor in angular velocity representation), in order to generate a co-rotating harmonic system. The other modulator can be configured to multiply the angle signal by the difference between 1 (simple rotational speed corresponding to the basic angular velocity) or 2π (simple basic angular velocity) minus the harmonic angular velocity, in order to generate a counter-rotating harmonic system.These modulation factors take into account that the space vector representation of the basic control loop already rotates at the basic angular velocity and is thus not stator-fixed, but rotor-fixed, and that the harmonic component rotates relative to the fundamental component and thus relative to the rotor, not relative to the stator. In other words, the angle signal is modulated to generate a co-rotating and a counter-rotating harmonic system that rotates relative to the rotor-fixed basic system at the harmonic angular frequency, or in the ratio of the harmonic angular frequency to the fundamental angular frequency. Rotation at the fundamental angular frequency already represents a simple rotation (at 1 or 2π, respectively), while the harmonic systems are based on the already rotating basic system.

[0038] A transformation or mapping can be connected downstream of the setpoint input, which converts the input values ​​into quantities in space vector representation, in particular into input controlled variables. This transformation can be referred to as a space vector transformation and can also include a conversion of the setpoint variables (speed, torque) into setpoint currents (in space vector representation or as parameters thereof). This results in the currents Id, Iq in space vector representation. The actuator of the basic control loop, in the form of an inverse space vector transformation, receives the output variable of the controller (after feeding in the counter signal in space vector representation, modulated with the harmonic angular frequency to match the fundamental-related space vector representation in the basic control loop). From this variable, the inverse space vector transformation of the basic control loop generates a (multi-phase) signal in the time domain to control the inverter.This signal is pulse width modulated and is input to a control input of the inverter.

[0039] According to one embodiment, the counter-signal generation can be designed for optimization according to three optimization objectives, ie (i) to minimize the harmonic component in the actual value, (ii) to minimize the harmonic component in the power supply signal and (iii) to reduce the harmonic component in the torque of the electric machine.

[0040] A further embodiment provides that the counter signal provided in space vector representation generates two mutually perpendicular voltage vectors, with a first of these vectors representing the magnetic flux density of the harmonic and a second of these vectors indicating the torque of the harmonic. Particularly when considering reluctance effects (or generally cross-dependencies between the vectors and the variables), the first of the vectors represents at least a portion of the flux density (and a portion of the torque), and the second vector represents a (different) portion of the torque (and a different portion of the flux density). Cross-dependencies may therefore exist between the vectors and the two variables; however, these do not preclude optimization / minimization or a sufficiently accurate representation of both variables.If these variables are split into two vectors with separate influences on the respective variables, depending on the harmonic, the second vector is adjusted to minimize the harmonic component in the torque. The first vector can be adjusted to minimize a reactive component that is included in the harmonic component in the actual value, or that corresponds to the harmonic component within the actual value. Both vectors can be adjusted to adjust or minimize / optimize the harmonic component in the torque, as well as to adjust or minimize / optimize the reactive component.

[0041] It can be provided that the counter-signal generation is configured to generate the counter-signal from the harmonic component by adjusting the two vectors of the space vector representation both to minimize the harmonic component in the actual value and / or in the power supply signal, as well as to reduce a harmonic component in the torque of the electric machine, and also to reduce the signal strength of the counter-signal. A further control objective can therefore be to minimize the signal strength of the counter-signal, or the counter-signal is generated with as low an amplitude as possible (without impairing the other control objectives).

[0042] The counter-signal generation is preferably configured to minimize the harmonic component in the actual value and / or in the power supply signal as a first secondary condition (or as a control objective) when generating the counter-signal. It can further be provided that the harmonic component in the torque of the electric machine is minimized as a second secondary condition (or as a further control objective). The first secondary condition and the second secondary condition can be incorporated into the generation of the counter-signal according to a respective, predetermined weighting. Thus, all or at least two of the aforementioned control objectives can be provided simultaneously (effectively) for configuring the counter-signal generation, wherein the aforementioned objectives are weighted relative to one another, in particular when pursuing one objective leads to a change in pursuing another objective.

[0043] The harmonic detection system can be configured to determine the harmonic component as a signal strength of the harmonic component along with the associated phase angle. In particular, the harmonic component can be determined together with a load angle and / or a power factor.

[0044] For harmonic detection, there are several ways to generate the harmonic component from the actual value or from the power supply signal (generally: power signal). For example, the harmonic component can be generated from the power signal (actual value or power supply signal, DC) using Fourier analysis, bandpass filtering, integration after modulation with at least one harmonic frequency, or by transforming the actual value (or the power supply signal or an AC component thereof) into a space vector representation related to a harmonic angular frequency. In general, the harmonic component of the actual value or the harmonic component of the power supply signal (or its AC component) can be determined, or both harmonic components can be determined.

[0045] The harmonic detection device is preferably configured to determine the harmonic component in the actual value (and / or in the power supply signal) as a discrete spectral representation of the harmonic component, which is assigned a frequency-specific phase position, as a representation of the harmonic component in the time domain, or as a space vector representation rotating at a harmonic angular frequency. The harmonic component is determined, in particular, together with phase information.

[0046] The actual value can be fed back as a multiphase signal in the time domain for deviation detection or in a space vector representation rotating at a harmonic angular frequency. The latter can also apply to the harmonic component in the power supply signal.

[0047] Embodiments of the control device have a counter-signal modulator configured to modulate at a harmonic angular frequency. The counter-signal generator can be connected via the counter-signal modulator to a point in the basic control loop between the actuator and the deviation detector. The modulator adjusts the counter-signal to the angular frequency (fundamental angular frequency), which forms the basis for signal representations in the basic control loop.

[0048] Further embodiments of the control device include a harmonic modulator. This is configured to modulate with a harmonic cyclic frequency. The harmonic detector is preferably connected via the harmonic modulator to a feedback deviation detector (SUB) provided in the feedback path, preferably between an actual value tap (measuring device) or a space vector transformation connected downstream thereof, and the deviation detector or the point at which the feedback of the basic control loop meets the reference variable (control path / forward path) or the control path. The feedback deviation detector is configured, in particular, to subtract the harmonic modulator signal from the feedback path signal or to determine the difference between these variables.

[0049] An electric vehicle traction drive can be equipped with an electric motor and an inverter, as well as with a control device as shown herein. The control device is connected, in particular, to the inverter (and possibly to the electric motor or an associated measuring device) in a control or signal-transmitting manner.

[0050] The space vector representations mentioned here, in particular the d,q representations, are obtained, for example, through a Park transformation. Furthermore, the space vector representation can be a representation of the positive and / or negative sequence system of symmetric components, which can be obtained, for example, through a Fortescue transformation. The space vector representation is preferably a rotor-related space vector representation. Alternatively, space vector representations that refer to the stator can also be used, such as space vector representations obtained using a Clarke transformation. Thus, the space vector representation can also be an α,β representation.

[0051] With regard to the objective of minimizing the harmonic content of the power supply signal (or in the actual value, especially if its harmonic content is reflected in the former signal), the following consideration can be provided: In simplified terms, the regulatory objective can be expressed as a condition by: P_VE=I_d'(ω_O)∗U_d(ω_G)+I_d'(ω_G)∗U_d(ω_O)+I_q'(ω_O)∗U_q(ω_G)+I_q'(ω_G)∗U_q(ω_O) with: - P_VE: Power of the harmonic component at the supply input of the inverter - I_d': first component of the space vector representation of the current I - I_q': second component of the space vector representation of the current I - U_d: first component of the space vector representation of the voltage U - U_q: second component of the space vector representation of the voltage Uwhere the current I indicates the current at the inverter control output or in the electrical machine, the voltage U indicates the voltage at the inverter control output or in the electrical machine, the suffix (ω_O) to a quantity indicates that the space vector representation of the quantity refers to the harmonic angular frequency, the suffix (ω_G) to a quantity indicates that the space vector representation of the quantity refers to the fundamental angular frequency and the designation of the variables corresponds to the reference symbol of the Fig. 1 corresponds.

[0052] The quantities P_VE, I_d', I_q', U_d', and / or U_q' can also be standardized quantities, for example, standardized to a total power, a rated current, an operating current (effective value), a rated voltage, or an operating voltage. The symbol * can represent a convolution if the spectral ranges of the quantities in question must be considered, or can represent a multiplication if the quantities in question are scalars, such as scalars in a space vector representation.

[0053] In simplified terms, the condition (control objective: minimizing the harmonic content in the power supply signal) can thus be represented by the (minimization of) the sum of all products of a space vector component of the current and a space vector component of the voltage, where the two factors of each product refer to space vector representations of different angular frequencies (fundamental angular frequency or harmonic angular frequency). Current and voltage refer to the inverter control output and the electric machine, respectively. The formula I_d*U_d + I_q*U_q = 0 can be provided as one of several control objectives or reduction conditions. Minimizing the d and q components of the harmonic content in the phase current (e.g., minimizing I_d'(ω_O) and I_q'(ω_O)) can also be provided as one of several control objectives.Furthermore, one of several control objectives can be to minimize the amplitude of the counter signal, for example by minimizing the signals U_q(O), U_d(O) or a combination thereof (for example their sum or the square root of the sum of the squares of these signals).

[0054] This applies in particular to a space vector representation of the counter signal, where the minimization of a moving average of the instantaneous power of the counter signal can also correspond to one of several control objectives.

[0055] The Fig. 1 serves to explain the control device described here and the control method implemented thereby using exemplary embodiments in a symbolic representation.

[0056] The Fig. Figure 1 shows a control device with a basic control loop (GRK) with a control path that includes a setpoint input (EG), a downstream deviation detector (RA), an actuator (IRT) (configured for pulse width modulation), and an inverter connected downstream of the actuator (IRT). The inverter controls an electrical machine.

[0057] A feedback path FB_G of the basic control loop feeds a current signal (I_U, I_V, I_W) from the inverter INV back to the deviation detector RA. The basic control loop GRK operates (partially) with signals in space vector representation, characterized by the suffixes d and q for the two vectors of the space vector representation. The space vector representation within the basic control loop GRK is based on a fundamental oscillation angular frequency ω_G. In particular, a (multiphase) signal representation in the time domain is used between the actuator IRT, which transforms a signal U_d, U_q into pulse-width-modulated signals in the time domain, and a space vector transformation RT_G within the feedback path FB_G, particularly with regard to pulse-width-modulated signals. Otherwise, a space vector representation is used in the basic control loop GRK (related to ω_G).

[0058] The Fig. 1 further shows a harmonic control loop ORK within the control device. Power signals such as the multiphase phase current I_U - I_W and / or a power supply signal such as a supply voltage U_HV or a supply current I_HV of the inverter INV serve as inputs. In particular, harmonic components (especially their signal strengths, possibly including phase information) of these power signals serve as inputs. Fig. 1 shows one of several possible configurations for determining the harmonic components, where in the Fig. 1 illustrates the exemplary determination of the harmonic components of the phase currents I_U - I_W in more detail. According to this variant, the power signal is represented using a space vector representation based on a harmonic angular frequency. Since the components that are not related to the harmonic angular frequency cancel each other out, the relevant harmonic component is obtained (already in space vector representation), see reference symbol RZO. The harmonic control loop ORK has a counter-signal generator GS_RZ connected downstream of the input. This is symbolically represented by a reduction condition RBF. Counter-signal generation GS_RZ, in addition to the reduction condition RBF (which aims to minimize the harmonic component in the power supply signal), also provides, as a further objective, the reduction of a harmonic component in the torque of the electric machine.For the latter goal, a calculation of the (harmonic component of the) torque of the electric machine EM is provided in the harmonic control loop ORK, or the harmonic component of the torque is already minimized within the counter-signal generation GS_RZ or within the reduction condition RBF. The torque (and its harmonic component) can be derived from the phase current (in the time domain or in space vector representation) and the magnetic properties of the electric machine (in particular its inductance), preferably taking into account the mass moment of inertia with respect to the harmonic component. In particular, the torque is derived from the phase current and the magnetic flux that is thereby generated in the electric machine. The counter-signal generation GS_RZ orThe reduction condition RBF can be considered a harmonic deviation detection, where the setpoint for the harmonics is zero or corresponds to a minimum value (minimized desired value), and the actual value of the harmonics corresponds to the harmonic component in the actual value of the basic control loop or in the power supply signal, or the harmonic component in the torque. Depending on the deviation (from zero or the minimum value), the counter signal is generated.

[0059] The input variable of the harmonic control loop ORK are the harmonic components in the actual value (current of the inverter / phase current of the electric machine) or in the power signal U_HV or I_HV. These harmonic components can be recorded from these input variables (which are present in the time domain) by space vector transformation RT_O, which relates to the harmonic angular frequency w_O (to be considered). Here, the 6th harmonic is considered as an example with ω_O = 6 * ω_G. The current angle θ(t) is taken into account. The space vector representation of the harmonic component includes both a space vector representation that rotates (co-rotates) with the fundamental frequency-related space vector representation (i.e., related to ω_G), and a space vector representation that rotates in the opposite direction (counter-rotates).To do justice to the existing reference to the fundamental harmonic representation, the space vector representation of the harmonic component results in a co-rotating component with an angular velocity of ω_G + ω_O (in the example for the 6th harmonic: 1 + 6 = 7) and a counter-rotating component with an angular velocity of ω_G - ω_O (in the example for the 6th harmonic: 1 - 6 = -5). This is represented by the harmonic modulations OM with *(-5) and with *(+7), with which ω_G is modulated, i.e. for the -5th order harmonic (-5th Od.) and for the 7th order harmonic (+7th Od.).

[0060] The harmonic component RZO generated in this way by space vector transformation RT_O is fed to the modulator MOD' after optional low-pass filtering TP and, above all, also fed as I_d(O) and I_q(O) (i.e. as space vector representation I_O_RZ) to the counter signal generator GS_RZ, which also works with signals in space vector representation. Instead of the harmonic detection based on space vector representation (reference symbol RT_O) described above, the harmonic component can also be determined from the actual value I_U - I_W and / or from the power supply signal (present in the time domain) by a harmonic detection OE that works in the time domain and / or in the frequency domain. This is shown with the dashed lines connected to the harmonic detection OE. The result is a representation of the harmonic component I_O, which is fed to the counter signal generator GS_RZ and, if necessary, converted into a space vector representation before feeding.The harmonic detection OE preferably also receives the angle θ(t) of the electrical machine EM in order to output the harmonic component in phase or angle related terms.

[0061] In addition to the harmonic component I_O_RZ, the counter signal generator GS_RZ also receives the actual value I_dq(G) and the control variable (output of the controller REG) U_dq(G). The counter signal generator GS_RZ operates in space vector representation with two components, denoted by d and q. The counter signal generator GS_RZ has a reduction condition RBF that refers to the two components of the space vector representation. This results in several degrees of freedom with regard to the control objectives, allowing multiple control objectives to be pursued simultaneously.

[0062] A first control objective or a first condition A of the reduction condition is to minimize the harmonic components in the actual value I_EM. This is Fig. 1 is represented by the equation I_d,q*U_d,q (O)= 0 (written out: I_d*U_d + I_q*U_q = 0), which symbolically provides a value of zero (or a minimum value) for the power I*U of the harmonic (O). A second control objective or a second condition B of the reduction condition provides for the reduction of a harmonic component T_O in the torque of the electric machine. This is symbolically represented by the equation T_O = min with "min" as the minimum value. A third control objective or a third condition C of the reduction condition provides for the reduction of a harmonic component in the power supply signal U_P (U_P stands for the power signal supplied to the inverter, i.e. the voltage, current or power signal resulting from U_HV and I_HV). This is symbolically represented by the equation U_P = min with "min" as the minimum value.The bundle of control objectives to be pursued can be objective A and objective B (without objective C), objective B and objective C (without objective A), or objective A, objective B and objective C. Objective A can be provided as the control objective, and objectives B and / or C can be provided as subordinate constraints. Objective C or objective B can also be provided as the control objective, and at least one of the remaining objectives can be provided as a constraint. A lookup table LT can be maintained to provide associated optimized values ​​(according to the reduction condition) for different operating points and different harmonic component strengths, which characterize the counter signals optimized (minimized) according to the reduction condition. The amplitude and / or phase position of the two components d, q of the space vector representation GS_ω_O (with U_q(O) and U_d(O) as components) can be used as identifiers.The space vector representation GS_ω_O represents the counter signal. The lookup table LT transmits the optimized values ​​QE to the counter signal generator GS_RZ. It is shown that one of the two vectors (q-vectors) can optionally be transmitted directly from the lookup table LT as part of the counter signal (without involving the counter signal generator GS_RZ), since the lookup table already contains minimized or optimized values. The signal U_q(O) is optimized according to at least one of the objectives A - C, so that it does not need to be subjected to the reduction condition RBF again. U_q(O) and U_d(O) can be provided separately according to several of the conditions A, B, C to generate an optimized counter signal. The two variables are not necessarily assigned to exactly one condition each; rather, one or more of the conditions can affect both variables (d, q components).When applying two of the objectives, there may be several numerical ranges of U_d(O), U_q(O) that are optimized according to these two objectives (A,B or A,C or B,C), so that when pursuing the third, remaining objective (C, B or A), one can select from these ranges and thus optimize the result. Therefore, with the two variables U_q(O) and U_d(O), more than two objectives can be applied in the optimization, especially if the objectives are weighted and incomplete minimizations (i.e. no global minimum) are also taken into account when achieving the objective. The term minimization is to be understood as meaning that a value is obtained that is smaller than a value that would result without a counter signal or without a harmonic control loop.

[0063] The counter signal (in space vector representation) output by the counter signal generator GS_RZ is fed to the input of the basic control loop via a modulator MOD. The modulator MOD modulates the counter signal GS_ω_O output by the counter signal generator GS_RZ (whose space vector representation is based on rotation at the harmonic angular velocity ω_O) with the harmonic angular velocity ω_O to account for the fact that the input is provided in the basic control loop GRK, whose variables at the input EGS are based on a space vector representation whose phasors rotate only at the fundamental angular velocity ω_G. In the example shown, this corresponds to the 6th harmonic, i.e., 6 * ω_G.Alternatively, the signal GS_ω_O can be transformed into a time-domain signal by inverse space vector transformation (not shown) to be superimposed on a control or positioning signal of the basic control loop that is present in the time domain, for example, at the feed EGS'. An optional limiter Lim can be provided between the counter-signal generator and the feed, for example, between the counter-signal generator GS_RZ and the modulator MOD (or also between the modulator MOD and the feed EGS).

[0064] In the basic control loop GRK, signals in space vector representation REF_RZ are used, which refer to the fundamental angular frequency ω_G, which corresponds to the electrical speed. This applies, for example, to the d and q components I_d and I_q, which represent a (multi-part, two-dimensional) setpoint variable. This variable is generated from the input variables EG using the setpoint variable generation SE, which, among other things, generates the variables I_d and I_q from the setpoint speed n and the setpoint torque Tq using space vector transformation RT. These represent the setpoint current in the electrical machine EM / the inverter INV in space vector representation - related to ω_G. Other setpoint parameters or input parameters (to be taken into account when generating I_d and I_q) can be the temperature Temp (of the electrical machine or the inverter) and / or the level of the supply voltage U_HV (= power supply signal) of the inverter INV.

[0065] Setpoint generation SE is followed by deviation detection RA. This is two-part, meaning it is provided once for each vector of the space vector representation. Deviation detection RA determines the deviation of the signal fed back from the feedback path FB_G to deviation detection RA (actual variable - feedback variable R) from the setpoint variables I_d and I_q. The feedback variable R is available in a space vector representation that refers to ω_G.

[0066] The deviation detection RA is followed by a (two-part) controller REG. The two parts REG', REG'' of the controller REG correspond to the two phasors of the space vector representation, i.e. the d and q components. The controllers REG output the manipulated variable U_dq(G). This is also available in a space vector representation that refers to the fundamental angular frequency. The counter signal GS is fed into the variable U_dq(G), which is output by the controller REG. For this purpose, an input EGS receives both the variable U_dq(G) output by the controller REG and the counter signal GS. In the example shown, both variables fed to the input EGS are available in a space vector representation that refers to the fundamental angular frequency ω_G. For this purpose, the counter signal was modulated by a modulator MOD with the harmonic frequency ω_O before the thus modulated signal was fed to the feed EGS.This results in the components U_d and U_q, which correspond to the signal U_dq(G) modulated by the counter signal GS. This diagram shows the counter signal being fed into the control variable, with these variables and the feed-in being based on a space vector representation. However, variables that exist in the time domain are also conceivable at this point in the control process. This also applies to the controllers and the deviation detection, as well as the setpoint variables.

[0067] The modulated space vector representation RZG of the (two-part) control variable U_d, U_q of the Fig. 1 is fed to an actuator IRT of the basic control loop. In addition to an inverse space vector transformation (which would be omitted if the variables were controlled in the time domain, as would the space vector transformation RT), the actuator generates a (multi-phase) pulse-width modulated signal PWM_U - PWM_W (reference symbol PS). When generating this signal, which is present in the time domain, the instantaneous angle θ(t) of the electric machine EM, which is detected by an angle sensor SN, is preferably taken into account. The angle sensor SN is connected to the shaft of the electric machine. The angle θ(t) can also be detected using the current of the electric machine EM, in addition to or alternatively to the angle sensor SN. The multi-phase pulse width signal PS is output, in particular, at a modulation output MA.If the counter signal is fed in in the time domain instead of in space vector representation, the counter signal can be taken into account when generating the pulse width signal PS, thus generating the counter signal in the time domain during pulse width modulation and generating it as a component of the pulse width signal PS. This is symbolically represented by the counter signal feed point EGS'.

[0068] The pulse width signal PS is applied to the inverter INV to control it. The inverter INV receives a supply signal from a battery Batt (high-voltage traction battery) in the form of the DC power voltage signal U_HV and the DC power current signal I_HV. The signals U_HV and I_HV can be referred to as the power supply signal. The power supply signal is essentially a DC signal, but contains (unwanted) AC components in the form of harmonics due to the operation of the inverter INV and the electric machine EM. The dot-dash line shows the tap of a signal that represents the power supply signal including the harmonic components. This signal is fed to a harmonic detector OE (which operates in the time domain) or RT_O (which operates in the space vector representation).

[0069] The inverter INV has an input I_IN for receiving the pulse width signal PS, a supply input for receiving the power supply signal U_HV, I_HV, and a multiphase output for connection to the electric machine EM. A tap or measuring device SM is provided to acquire the actual value I_EM of the basic control loop. The actual value I_EM is available in the time domain.

[0070] In the Fig. 1, the actual value I_EM is fed to a space vector transformation RT_G of the feedback loop R or the feedback path FB_G. If the feedback loop is operated in the time domain, this transformation RT_G is omitted. The space vector transformation RT_G generates the space vector representation I_G_RZ of the electrical machine's current, which is also represented by the components I_d' and I_q'. This current is represented in space vector representation, relative to the fundamental angular frequency G, and corresponds to I_dq(G), i.e., a representation in d,q components.

[0071] To prevent the two-part controller REG from detecting the counter signal G as a deviation and compensating it through control, a subtraction element SUB is connected downstream of the space vector transformation RT_G within the feedback path FB_G of the basic control loop. A modulated space vector representation of the harmonic component ω_O is fed there. For this purpose, the detected harmonic component RZO is fed to a modulator MOD', which modulates it with the harmonic angular frequency ω_O (signal waveform: cos(ω_O)). In the example shown, this is 6 times the fundamental angular frequency ω_G as an example of the 6th harmonic to be compensated. The modulated harmonic component fed to the subtraction element SUB is subtracted from the signals I_d' and I_q' by the subtraction element SUB. The resulting feedback variable R of the basic control loop is fed to the deviation detection element RA.Since the control deviation forms the input variable for the two-part controller REG, it will not react to the harmonic components I_O_RZ, since, as mentioned, these have already been removed in advance by the subtraction from the feedback R of the basic control loop.

[0072] The following relationship can be used to calculate the harmonic component in the torque (torque ripple): idqtudq=idqtRdqidq+idqtωelJ[Ldqidq+ΨPM]+idqtωel[dLdqdφidq+dΨPMdφ] +idqt[ddtidq]Ldq

[0073] The term to the left of the equal sign in equation 2 represents P el , ie the electrical power due to the current i dq and the tension u dq (instantaneous values ​​in each case).

[0074] The first summand in equation 2 corresponds to_P j , ie the current ohmic heat loss through resistance R dq , and the second term in equation 2 corresponds to P m, ie mechanical power, where ω el the electrical angular velocity, L dq the inductance and Ψ the magnetic flux.

[0075] The third summand of equation 2 corresponds to a varying mechanical power that is related to magnetic quantities (L dq - inductance, Ψ - magnetic flux), which vary in the direction of rotation, ie with the angle φ (electrical angle, linked to rotor angle θ).

[0076] The fourth term of equation 2 corresponds to P mag , i.e., the magnetic power required for the build-up and dissipation of magnetic energy in the electrical machine. The current i, the resistance R, and the inductance L are each represented by d,q components; for multiphase machines, additional independent components are added. The flux, designated "PM," refers to a flux offset due to the permanent magnet.

[0077] The superscript t stands for transpose, so that for a matrix or vector with a superscript t, the result is the transposed matrix.

[0078] In equation 2, J represents a fixed transformation matrix resulting from the Park transformation: J=[0−100⋯1000⋯000−3⋯0030⋯⋮⋮⋮⋮⋱]

[0079] Numerous sizes in Fig. 1 are shown in space vector representation, this is generally indicated by the suffix RZ in the reference symbols, and specifically for individual quantities with the suffix “_d”, “_q” as d and q components in the respective quantity in space vector representation.

[0080] In the illustrated embodiment, the space vector representation is rotor-fixed, i.e., it represents a system or coordinate axes that are fixed on the rotor (or are fixedly linked to the rotation system generated by the stator) and thus assume rotation at a specific angular frequency. This angular frequency corresponds to the fundamental angular frequency ω_G, particularly within the fundamental control loop GRK. The space vector representations in the harmonic control loop ORK are based on the harmonic angular frequency ω_O, unless otherwise stated.

[0081] The following reference symbols are used in the figure: GS: Opposite signal GSS: Signal strength of the opposing signal R: Repatriation FB_O: Counter signal feedback EG: Input variables, setpoint input RA: Control deviation, deviation recording RS: Control variable control REG: linear regulator GSE: Counter signal feed EG_T: Transformation of the input variable into space vector representation Additional _RZ: Space vector representation PS: Pulse width signal MA: Modulation output RZG: Space vector quantities (suppressed) IRT: Inverse space vector transformation RT_O: Harmonic space vector transformation / harmonic detection INV: Inverter I_IN: Inverter control input I_OUT: Inverter control output I_EM: Output signal of the inverter SN: Sensor θ: rotor angle, θ(t): time signal of the rotor angle ω_G: fundamental angular frequency ω_O: harmonic angular frequency OM: Harmonic Modulator TP: Low pass I_d(O), I_q(O): complex d, q - components of the harmonic I_O_RZ - complex space vector representation of the harmonic EGS: injection of the counter signal (GS) Lim: Limiter GS_RZ : Counter signal calculation in space vector representation MOD, MOD': Modulator RT: Transformation into a space vector representation SM: Current measuring device (for measuring the operating currents of the EM) GRK: Fundamental oscillation control loop ORK: Harmonic control loop LT: Lookup table

[0082] The components functionally designated here are preferably designed as device elements.

Claims

[1] Control device for controlling an electric drive comprising an inverter (INV) and a multiphase electric machine (EM) connected thereto, wherein the control device comprises a basic control loop (GRK) with a deviation detection system (RA), an actuator (IRT) and a controlled system (RS) comprising the inverter (INV) and the electric machine (EM), wherein in the feedback loop (FB_G) of the basic control loop (GRK) the current of the multiphase current (I_U, I_V, I_W) flowing between the inverter (INV) and the electric machine (EM) is fed back to the deviation detection system (RA) as an actual value (I_EM), wherein the basic control loop (GRK) is referenced to a fundamental angular frequency (ω_G) and wherein the control device comprises a harmonic detection system (OE, RT_O) which is connected downstream of the feedback loop (FB_G) of the basic control loop (GRK) and is configured to detect the harmonic component (I_O) in a power supply signal (U_P),to determine the inverter (INV), and has a counter-signal generation (GS_RZ) which is connected downstream of the harmonic detection (OE) and is configured to generate a counter-signal (GS) which is fed into the control variable (RZG) of the actuator (IRT) via a feed (EGS) connected upstream of the actuator (IRT), wherein the counter-signal generation (GS_RZ) is configured to generate the counter-signal (GS) provided in a space vector representation from the harmonic component (I_O) by adjusting the two vectors of the space vector representation both to minimize the harmonic component (I_O) in the power supply signal (U_HV) and to reduce a harmonic component in the torque of the electric machine (EM), wherein the counter-signal (GS) provided in the space vector representation generates two mutually perpendicular voltage vectors,wherein a first of these pointers (U_q(O)) represents at least a component of the magnetic flux density of the harmonic and a second of these pointers (U_d(O)) characterizes at least a component of the torque of the harmonic, wherein the first and the second pointers are adjusted according to a minimization of the harmonic component in the torque according to a minimization of a reactive component, wherein the reactive component is provided in the harmonic component (I_O) in the actual value (I_EM), or the reactive component corresponds to the harmonic component (I_O) in the actual value (I_EM). [2] Control device according to claim 1, wherein the counter-signal generation (GS_RZ) is configured to adjust the two pointers of the space pointer representation to minimize the harmonic content (I_O) in the actual value (I_EM), to minimize the harmonic content in the power supply signal (U_HV) and to reduce the harmonic content in the torque of the electric machine (EM). [3] Control device according to claim 1 or 2, wherein the counter-signal generation (GS_RZ) is configured to generate the counter-signal (GS) from the harmonic component (I_O) by adjusting the two pointers of the space pointer representation both to minimize the harmonic component (I_O) in the power supply signal (U_HV) and to reduce a harmonic component in the torque of the electric machine (EM) and also to reduce the signal strength of the counter-signal (GS). [4] Control device according to one of the preceding claims, wherein the counter-signal generation (GS_RZ) is configured to minimize the harmonic component (I_O) in the power supply signal (U_HV) as a first constraint during the generation of the counter-signal (GS) and to minimize the harmonic component in the torque of the electric machine as a second constraint, wherein the first constraint and the second constraint are incorporated into the generation of the counter-signal (GS) according to a respective, predetermined weighting. [5] Control device according to one of the preceding claims, wherein the harmonic detection (OE, RT_O) is configured to determine the harmonic component (I_O; I_d(O), I_q(O)) as a signal strength of the harmonic component (I_O; I_d(O), I_q(O)) together with the associated phase angle. [6] Control device according to one of the preceding claims, wherein the harmonic detection (OE, RT_O) is configured to generate the harmonic component (I_O; I_d(O), I_q(O)) from the power supply signal (U_HV) by Fourier analysis, bandpass filtering, integration after modulation with at least one harmonic frequency, or by transforming the actual value (I_EM) into a space vector representation that rotates with a harmonic angular frequency. [7] Control device according to one of the preceding claims, wherein the harmonic detection (OE) is configured to determine the harmonic component (I_O; I_d(O), I_q(O)) in the actual value (I_EM) as a discrete spectral representation of the harmonic component (I_O; I_d(O), I_q(O)) to which a frequency-specific phase is assigned, as a representation of the harmonic component (I_O; I_d(O), I_q(O)) in the time domain, or as a space vector representation that rotates with a harmonic angular frequency (ω_O). [8] Control device according to one of the preceding claims, wherein the actual value (I_EM) is fed back as a multiphase signal in the time domain for deviation detection (RA) or is fed back in a space vector representation which rotates with a harmonic angular frequency (ω_O). [9] Control device according to one of the preceding claims, comprising a counter-signal modulator (MOD) configured to modulate with a harmonic angular frequency (ω_O), wherein the counter-signal generation (GS_RZ) is connected via the counter-signal modulator (MOD) to a point of the basic control loop (GRK) between the actuator (IRT) and the deviation detection (RA). [10] Control device according to one of the preceding claims, comprising a harmonic modulator (MOD') configured to modulate with a harmonic angular frequency (ω_O), wherein the harmonic detection (OE, RT_O) is connected via the harmonic modulator (MOD') to a feedback deviation detection (SUB) which is positioned between the feedback path (FB_G), wherein the feedback deviation detection (SUB) is configured to subtract the signal of the harmonic modulator (MOD') from the signal of the feedback path (FB_G). [11] Electric vehicle traction drive comprising an electric machine (EM) and an inverter (INV) and a control device according to one of the preceding claims, wherein the control device is connected to the inverter and the electric machine (EM).

Images