Method for determining current-dependent and / or angle-dependent parameters of an electrical machine and frequency converter

Abstract

Method for determining current-dependent and / or rotation-angle-dependent parameters of an electrical machine (1) for implementation by means of an inverter (8), comprising the steps: a) Setting a rotational angle position of a rotor of the electric machine (1) and subsequently locking the rotor, b) Forming a longitudinal flow setpoint by adding a longitudinal flow setpoint operating point ( I d ,0 k ) and a longitudinal current setpoint alternating signal ( I d , HF k ) , where the longitudinal current setpoint alternating signal ( I d , HF k ) periodically changed with a longitudinal flow setpoint frequency, and / or forming a transverse flow setpoint by adding a transverse flow setpoint operating point ( I q ,0 k ) and a cross-current setpoint alternating signal ( I q , HF k ) , where the cross-flow setpoint alternating signal ( I q , HF k ) periodically changed with a cross-current setpoint frequency, c) Rules of a longitudinal flow (i d ) on the longitudinal flow setpoint and / or control of a transverse flow (i q ) to the cross-current setpoint, whereby control variables in the form of a longitudinal control voltage are used for regulation. (udk) and / or a transverse control voltage ( uqk ) generated and impressed into motor phases of the electric machine (1), and d) Measuring phase currents (i u,ist , i v,ist , i w,ist ) of the electric machine (1) and determining an emerging longitudinal current (i d, ist ) and / or an emerging crossflow (i q, ist ) from the measured phase currents (i u,ist , i v,ist , i w,ist ), characterized by the steps: e) Applying a discrete Fourier transform (DFT) algorithm or a Görtzel algorithm to the longitudinal control voltage (udk) and on the determined longitudinal flow (i d,ist ) to generate longitudinal control voltage coefficients and longitudinal current coefficients and / or to apply the DFT algorithm or the Görtzel algorithm to the transverse control voltage ( uqk ) and on the determined cross-flow (i q,ist ) for generating transverse control voltage coefficients and transverse current coefficients, and f) Calculating the parameters as a function of the longitudinal control voltage coefficients and the longitudinal current coefficients and / or the transverse control voltage coefficients and the transverse current coefficients, wherein the calculation of the parameters is carried out as a function of a dead time of the inverter (8) and / or as a function of signal propagation times of filters used to measure the phase currents.

Description

[0001] The invention relates to a method for determining current-dependent and / or rotation-angle-dependent parameters of an electrical machine, in particular carried out by means of a frequency converter, and to a frequency converter.

[0002] Electric machines (electric motors) in the form of permanent magnet synchronous machines (PMSM) and synchronous reluctance machines (SynRM) offer significant advantages over the widely used asynchronous and induction machines with regard to their energy efficiency, particularly in the partial load and partial speed ranges, as well as power density and speed synchronization. However, cost-effective, sensorless operation of these synchronous electric motors requires a very precise understanding of their characteristics in the form of parameter-based equivalent circuit data. Due to the saturation-dependent properties of electric motors, parameters such as inductance typically change with the current or depending on the rotor position.

[0003] The publication by J. Kiel et al.: Determination of operating point-dependent machine parameters for synchronous machines, SPS / IPC / DRIVES Conference Proceedings 2000, pp. 814 to 821 shows a method according to the preamble of claim 1.

[0004] The publication by J. Weigel and P. Mutschler: “Acquisition of a permanent magnet linear synchronous motor's electrical parameters and dead beat current control considering saturation,” IEEE International Electric Machines and Drives Conference, 2003. IEMDC03., 2003, pp. 791-797 vol.2 discloses a method for determining saturation-dependent inductances of a synchronous machine.

[0005] The invention is based on the objective of providing a method for determining or identifying current-dependent and / or angle-dependent characteristic parameters of an electrical machine and a frequency converter.

[0006] The invention solves this problem by means of a method according to claim 1 and a frequency converter according to claim 11.

[0007] The method is used to determine current-dependent and / or rotation angle-dependent parameters of an electrical machine.

[0008] The parameters can be used, for example, for model-based, sensorless control of PMSM or SynRM, as well as for efficiency- or current-optimized control (Maximum Torque per Ampere Control or MTPA), or for testing the motor's characteristics. Furthermore, the information can be used for improved current setpoint feedforward control.

[0009] First, the rotational angle of the rotor of the electric machine is set. Once the desired rotational angle is reached, the rotor is blocked, particularly mechanically, so that it can no longer change its rotational angle. This blocking can be achieved, for example, with a service or holding brake.

[0010] Subsequently, a longitudinal current setpoint is generated by adding a longitudinal current setpoint operating point and a longitudinal current setpoint AC signal, whereby the longitudinal current setpoint AC signal changes periodically at a longitudinal current setpoint frequency, and the longitudinal current setpoint operating point remains constant at least for certain durations. Alternatively or additionally, a transverse current setpoint is generated by adding a transverse current setpoint operating point and a transverse current setpoint AC signal, whereby the transverse current setpoint AC signal changes periodically at a transverse current setpoint frequency, and the transverse current setpoint operating point remains constant at least for certain durations. The periodic AC signals, which can be sinusoidal, for example, serve to identify the differential impedance or inductance at the operating point defined by the constant current components.

[0011] Phase currents (currents in the motor phases) of the electric machine are now measured, and the resulting longitudinal current and / or transverse current is conventionally determined from the measured phase currents and a measured or estimated rotor flux position. Reference is made to the relevant technical literature in this regard.

[0012] The longitudinal flow i d The system is now conventionally controlled to the longitudinal flow setpoint and / or the transverse flow i q The current is conventionally controlled to the transverse current setpoint, whereby control variables in the form of a longitudinal control voltage and / or a transverse control voltage are conventionally generated and applied to the motor phases of the electric machine. Reference is made to the relevant technical literature in this regard.

[0013] In order to achieve good control behavior of the current controller with respect to both the operating point and the RF component, as well as good decoupling between the longitudinal and transverse branches, the use of a PI-P resonant controller is suitable for both the transverse and longitudinal current.

[0014] The applied longitudinal and transverse control voltages and the measured longitudinal and transverse currents can be used to determine the differential impedance of the machine at the set operating point. If the control voltages or the measured currents contain interfering noise or other interfering or superimposed signal components, a filter (e.g., a second-order bandpass filter) can be used for filtering. It is important that the filter is applied to all control voltages and longitudinal and transverse currents included in the calculation, so that the phase shift and amplitude attenuation caused by the filter occur equally for all quantities.

[0015] A discrete Fourier transform (DFT) algorithm is applied to both the unfiltered and filtered control voltages and currents, respectively. The result is the complex Fourier coefficients for longitudinal and transverse current and longitudinal and transverse control voltage as a function of frequency. If an RF signal with a fixed frequency (e.g., a sinusoidal test signal) is used for excitation, the discrete Fourier transform can be replaced by the significantly simpler Goertzel algorithm, which can calculate the Fourier coefficients for current and voltage for the desired excitation frequency with considerably less computational time and memory requirements.

[0016] Finally, the characteristic parameters are calculated as a function of the longitudinal control voltage coefficients and the longitudinal current coefficients and / or the transverse control voltage coefficients and the transverse current coefficients. These characteristic parameters can include the impedance Z at the operating point. From the impedance, the ohmic resistance can be determined by separately considering the imaginary and real components, and, taking the excitation frequency into account, the differential inductances Ldd, Lqq, Ldq, and Lqd can be derived.

[0017] In step b) the longitudinal flow setpoint operating point can be changed and / or the transverse flow setpoint operating point can be changed, with steps c) to f) being repeated when the longitudinal flow setpoint operating point and / or the transverse flow setpoint operating point is changed.

[0018] Controlling the longitudinal flow i d on the longitudinal flow setpoint and / or the control of the transverse flow i qThe adjustment of the transverse current setpoint can be performed based on control parameters for the current controllers, which depend on the longitudinal current setpoint operating point and / or the transverse current setpoint operating point. If, for example, a PL or a PI-P resonant controller is used to control the current, these controllers can be adjusted using tuning methods known from the literature (e.g., magnitude optimum). The necessary parameters (inductances) for this method can be obtained, for example, from an operating point that has already been identified and which differs only slightly from the operating point currently being identified. In this way, stable and dynamic behavior of the current controllers can be ensured, for example, even with SynRM, whose differential inductance can change by a factor of 3 depending on the current.

[0019] The longitudinal flow setpoint operating point and / or the transverse flow setpoint operating point can be continuously changed in such a way that a change in the longitudinal flow and / or the transverse flow does not exceed a predefinable threshold. The thresholds for positive and negative longitudinal and transverse flows thus limit the field of operating points or define the search space to be identified.

[0020] The calculation of the characteristic parameters depends on the inverter's dead time and / or on the signal propagation times of filters used to measure the phase currents. The phase shift on the voltage signal resulting from the dead time can, for example, be used to calculate the impedance, thus avoiding incorrect calculations of the imaginary and real parts of the impedance.

[0021] In step a), the rotation angle position can be changed by releasing the blockage, and then the rotor can be blocked again with the rotation angle position changed, whereby steps b) to f) are repeated with the rotation angle position changed in order to determine the characteristic values ​​as a function of the rotation angle.

[0022] The current-dependent and / or angle-dependent parameters can be selected from the following set of parameters: differential (derived with respect to current) longitudinal inductance(s) L dd , differential (derived with respect to the current) transverse inductance(s) L qq , Cross-saturation L dq and L qd , and resistances.

[0023] Furthermore, impedances composed of inductance and resistance can also be used directly. Ldd(id,iq,φ)=∂ψd(id,iq,φ)∂Id, Ldq(id,iq,φ)=∂ψd(id,iq,φ)∂Iq Lqd(id,iq,φ)=∂ψq(id,iq,φ)∂Id, Lqq(id,iq,φ)=∂ψq(id,iq,φ)∂Iq.

[0024] For better understanding, the stress equations of a PMSM and a SynRM are shown in simplified form. For SynRM, it should be noted that the permanent flux ψ PM zero. Ud=Rs⋅Id+Ldd⋅dIddt+Ldq⋅dIqdt−ωel⋅Lsq⋅Iq Uq=Rs⋅Iq+Lqd⋅dIddt+Lqq⋅dIqdt−ωel⋅Lsd⋅Id+ωel⋅ψPM

[0025] Current-dependent and / or location-dependent flux linkages in the longitudinal and transverse directions Psi_d, Psi_q can be calculated from the calculated differential longitudinal and transverse inductances Ldd, Lqq, as well as the calculated cross-saturation. For permanent magnet synchronous machines, this can be done particularly by considering the longitudinal flux linkage caused by the magnets. ψd(id,iq)=∫0Id,maxLdd(id,iq)did+ψPM ψq(id,iq)=∫0Iq,maxLqq(id,iq)diq

[0026] Current-dependent and / or location-dependent absolute inductances in longitudinal and transverse directions can be calculated from the flux linkages in longitudinal and transverse directions as well as from the associated longitudinal and transverse currents. Lsd(id,iq)=ψd(id,iq)id, Lsq(id,iq)=ψq(id,iq)iq

[0027] The longitudinal currents and / or transverse currents obtained from the measurement of the phase currents can be filtered, for example by bandpass filtering.

[0028] The electrical machine can be a synchronous machine or a reluctance machine. The method can also be used for induction machines.

[0029] The frequency converter has a control unit, for example in the form of a microprocessor and associated software, which is designed to carry out the aforementioned procedure. It is understood that the frequency converter has suitable measuring instruments for measuring the required quantities.

[0030] The basic idea of ​​the method is to determine the characteristic or equivalent circuit parameters for each relevant operating point. For this purpose, different operating points, characterized by a constant longitudinal and transverse current, are approached using a SynRM or PMSM that is braked at a defined angular position. By injecting a test signal in the longitudinal and / or transverse direction, the differential impedances and inductances can then be determined for each operating point. Further mathematical calculations can then be used to derive the absolute inductances and fluxes from these differential values.

[0031] The invention is described in detail below with reference to the drawings. These show: Fig. 1 a basic schematic structure diagram of a method according to the invention for determining current-dependent and / or angle-dependent characteristic parameters of an electrical machine, Fig. 2 a schematic block diagram of a structure of the method according to the invention in greater detail for determining the current-dependent and / or angle-dependent characteristic values, Fig. 3. A representation of the principle of alternating excitation for a working point, Fig. 4 a time course of a longitudinal flow setpoint and a transverse flow setpoint, Fig. 5 a time course of a state of a state control system in conjunction with an associated time course of a longitudinal flow setpoint and a transverse flow setpoint, and Fig. 6 a part of the in Fig. 2 shown block diagram in greater detail.

[0032] The method according to the invention enables the identification of the current-dependent equivalent circuit data or characteristic values ​​of a permanent magnet synchronous machine, a synchronous reluctance machine, and / or an induction machine. The method can be carried out without a position sensor. However, it is necessary to brake the machine during the method, at least for certain periods of time.

[0033] The method is based on the alternating injection of RF signals. Signal injection can be performed using RF voltage signals or by injecting an RF current into a field-oriented coordinate system. The following description is based on the injection of the RF current. The measurement is performed with the rotor locked, and its position is either known or estimated. The rotor position or angle of rotation can be identified using established methods. Iron losses only become significant at higher rotational speeds and can therefore be neglected when measuring inductance with the machine at rest.

[0034] Fig. Figure 1 shows, as an overview, a structure diagram of a method according to the invention for determining current-dependent and / or rotation-angle-dependent characteristic values ​​in the form of differential longitudinal inductances L. dd , differential transverse inductances L qq , Cross-saturations Ldq and L qd and optionally resistances or impedances.

[0035] Referring to Fig. In an optional input step I1, the existing controllers are first parameterized using design data. Furthermore, the operating points to be measured are defined in I1.

[0036] In step S1, the rotational angle position of the rotor of the electric machine is then set, i.e., the rotor is aligned.

[0037] In the following input step I2, the rotor of the electric machine is blocked so that the rotational angle of the rotor can no longer change.

[0038] In step S2, the winding resistance and inverter characteristic curve are determined. Furthermore, the inductances are roughly determined using simple test signals (e.g., step responses). The information obtained in this way is used to design the frequency of the subsequent AC signal and to preset the control parameters of the current controller.

[0039] In step S3, the actual procedure for identifying the differential inductances, resistances, and inductances takes place.

[0040] In step S4, the flux linkages Psi_d and Psi_q as well as the absolute inductances are determined from the differential inductances.

[0041] In step S5, a chain of permanent magnet flux is determined taking into account a voltage constant ke, which is entered in an input step I3.

[0042] The determined parameters are stored in a database 14.

[0043] Fig. Figure 2 shows a detailed block diagram of a structure of the method according to the invention.

[0044] A brake 2 is used to block an electric machine 1 in the form of a synchronous machine, a reluctance machine or an induction machine.

[0045] A state controller 3 is used to generate a state variable k. The state controller 3, in conjunction with a setpoint generator 18, is used to specify a longitudinal current setpoint operating point. ID, 0k and a cross-flow setpoint operating point Iq,0k, each depending on the state k. The state k can take on integer values ​​between 0 and 2.

[0046] An RF generator 4 receives the state k and, after the steady-state operating point has been reached and the steady-state currents have settled, generates either a longitudinal current setpoint alternating signal depending on this. id,HFk or a cross-current setpoint alternating signal iq,HFk.

[0047] A current controller 5 receives the longitudinal current setpoint operating point. ID, 0k, the cross-flow setpoint operating point Iq,0k, the longitudinal current setpoint alternating signal id,HFk, the cross-flow setpoint alternating signal iq,HFk and the state k. In current control 5, a longitudinal current setpoint is formed by adding the longitudinal current setpoint operating point and the longitudinal current setpoint AC signal, and a transverse current setpoint is formed by adding the transverse current setpoint operating point and the transverse current setpoint AC signal. For clarification, see also Fig. 6. The current controller can be implemented as, for example, a PL or PI-P resonant controller and its controller parameters can be adjusted from operating point to operating point.

[0048] The current control 5 regulates the longitudinal current i d on the longitudinal flow setpoint and regulates the transverse flow i q on the cross-current setpoint, whereby control variables in the form of a longitudinal control voltage are used for regulation. udk and a transverse control voltage uqk The voltage is generated and impressed into the motor phases of the electric machine 1. For this purpose, a conventional transformation of the field-oriented control voltage quantities (dq system) into the stator-fixed α,β system or into the phase-voltage-based uvw system is performed (for example, to the inverse Clarke transform 6). The voltage can be impressed by means of a pulse-width modulation (PWM) generation unit 7 and a three-phase H-bridge inverter 8. Elements 6, 7, and 8 are conventional elements that are used, for example, in so-called field-oriented control or vector control systems. Reference is made to the relevant technical literature in this regard.

[0049] Referring to Fig. 6, phase currents i u,ist , i v,ist , i w,ist of the electric machine 1 using conventional measuring instruments (not shown), based on the measured phase currents i u,ist , iv,ist , i w,ist by means of a conventional uvw-dq converter 9 (for example, for the Clarke-Parks transformation) a set longitudinal current i d,ist and an emerging cross-flow i q,ist from the measured phase currents i u,ist , i v,ist , i w,ist to be determined. The phase currents i u,ist , i v,ist , i w,ist Measurements can be taken with either 3 or 2 measuring elements, whereby in the case of 2 measuring elements the third phase current must be calculated from the other two phase currents. Reference is made to the relevant technical literature in this regard. The currents i d,ist and i q,ist are provided as actual values ​​for the current control 5.

[0050] The current control 5 can have a PI controller and a P-resonance controller (PR controller) connected in parallel to the PI controller.

[0051] Referring again to Fig. 2, an identification unit 10 serves to determine the characteristic values ​​based on the determined longitudinal flow i d , the determined cross-flow i q , the longitudinal control voltage udk, , the transverse control voltage uqk and the state k.

[0052] A switch or multiplexer 15 is connected to the series current i d , the crossflow i q and is subjected to state k and, depending on state k, either gives the longitudinal current i d or the crossflow i q to an optional filter 11, which can be implemented, for example, in the form of a bandpass filter. The filter 11 (or corresponding further, parallel instances of the filter 11) is / are further connected to the longitudinal control voltage. udk and the transverse control voltage uqk imposed.

[0053] A DFT unit 12 applies a discrete Fourier transform (DFT) algorithm or a Görtzel algorithm to the longitudinal control voltage. udk, and / or the transverse control voltage uqk and / or the longitudinal flow i d and / or the crossflow i q for generating longitudinal control voltage coefficients, transverse control voltage coefficients, longitudinal current coefficients or transverse current coefficients.

[0054] A calculation unit 13 then calculates the (impedance) parameters or differential inductances as a function of the longitudinal control voltage coefficients and the longitudinal current coefficients, as well as the transverse control voltage coefficients and the transverse current coefficients. For this purpose, the calculation unit 13 can divide the longitudinal current coefficients by the longitudinal control voltage coefficients and the transverse current coefficients by the transverse control voltage coefficients. From the differential inductances calculated in this way, the flux linkages and the absolute fluxes can be calculated. During the calculation, dead times of the inverter as well as different filter times, e.g., during current measurement, can be taken into account and compensated for.

[0055] The calculated parameters are stored in database 14.

[0056] Furthermore, the determined parameters can be used to adjust the current control for one of the following operating points. For this purpose, adjustment of controller parameters 16 and 17 is provided.

[0057] Furthermore, for normal operation a setpoint generator 19, a downstream position controller 20 and a downstream speed control 21 are provided, wherein an output of the setpoint generation 18 and an output of the speed control 21 are led to a changeover switch 22, which, depending on a braking condition, either the output of the setpoint generation 18 or the output of the speed control 21 is given to the current control 5.

[0058] Fig. Figure 3 shows a representation of the principle of alternating excitation for a working point, which is explained below with reference to the Fig. 4 and Fig. 5 will be explained in more detail.

[0059] Fig. Figure 4 shows a time course of the longitudinal current setpoint (or longitudinal current) and the transverse current setpoint (or transverse current). The control of the operating points is optimized such that there is only a small difference between either the longitudinal current setpoint or the transverse current setpoint between two consecutive operating points. This advantageously implemented sequence of operating points makes it possible to use the measured impedances of the previous operating point for adaptively adjusting the controller parameters for measuring the next operating point. Thus, even machines with extreme saturation behavior and highly variable inductances can be identified stably, robustly, and reliably. One possible implementation involves initially keeping the longitudinal current setpoint constant at the operating points and varying the transverse current setpoint from 0 to the maximum transverse current setpoint to be identified. After reaching the maximum,The longitudinal current setpoint is increased based on the cross-flow setpoint, and the operating points for this longitudinal current setpoint in the desired quadrant can then be approached by stepwise decrementing the cross-flow setpoint from the maximum setpoint to 0. By cleverly combining these measurements, all four quadrants of a motor can be measured in this way.

[0060] The steps described above for determining the key parameters are repeated for the changed operating points.

[0061] Furthermore, the rotation angle can be changed and then the rotor can be blocked with the rotation angle changed, whereby the steps described above for determining the characteristic values ​​with the rotation angle changed are repeated.

[0062] Fig. Figure 5 shows a time course of the state k of the state control 3 in conjunction with an associated time course of the longitudinal current or longitudinal current setpoint and the transverse current or transverse current setpoint.

[0063] The state k initially has a value of 0, then changes its value to 1, then to 2, and returns to 0. This cycle is repeated continuously.

[0064] During the first cycle, both the longitudinal flow setpoint and the transverse flow setpoint remain constant. In this state, the new longitudinal and transverse flow setpoints of the operating point are approached. The system waits for the controllers to settle and for steady-state behavior to avoid an unwanted superposition of the operating point approach process and the actual identification (k=1, k=2).

[0065] In the state k = 0, neither the longitudinal current setpoint alternating signal nor the transverse current setpoint alternating signal is generated, so that the longitudinal current setpoint and the transverse current setpoint also remain constant.

[0066] In state k = 1, only the longitudinal current setpoint AC signal is generated as a sinusoidal signal with a longitudinal current setpoint frequency, so that the longitudinal current setpoint changes sinusoidally around its operating point with the longitudinal current setpoint frequency, as shown. The transverse current setpoint remains constant. In state k = 1, longitudinal inductances Ldd and cross-saturations Ldq are calculated. Switch 15 ( Fig. 2) and thus calculation method 10 are only activated after the alternating signal has been present for several periods and the controllers have been able to settle. The switch is then closed for an integer number of alternating signal periods.

[0067] In state k = 2, only the transverse current setpoint AC signal is generated as a sinusoidal signal with a transverse current setpoint frequency, so that the transverse current setpoint changes sinusoidally around its operating point with the transverse current setpoint frequency as shown. The longitudinal current setpoint remains constant. In state k = 2, transverse inductances Lqq and cross-saturations Lqd are calculated. Switch 15 ( Fig. 2) and thus calculation method 10 are only activated after the alternating signal has been present for several periods and the controllers have been able to settle. The switch is then closed for an integer number of alternating signal periods.

[0068] The longitudinal current setpoint frequency and the transverse current setpoint frequency can be identical. The frequency for the procedure must be selected depending on the motor being identified and the available control voltage. Depending on the motor, it can range between 100 and 600 Hz. The amplitude of the current signal must be determined depending on the resolution of the current sensing and the current range to be measured. As a rule, the amplitude should be between 5% and 15% of the machine's rated current.

[0069] In the following cycle, as in Fig. Figure 5 shows that the cross-flow operating point is increased, while the longitudinal-flow operating point remains constant. Otherwise, the second cycle is identical to the first. Over a longer period, this ultimately results in the following: Fig. 4 shows the time profile of the current setpoint values.

[0070] According to the invention, longitudinal and transverse currents are controlled by a parallel-connected PI controller and PR controller (proportional resonant controller). The corresponding current setpoints are generated by a state control 3 in conjunction with a setpoint generation unit 18. The corresponding setpoint voltages from the output of the controller 5 and the measured currents are analyzed with the Goertzel algorithm with respect to the longitudinal current setpoint frequency and the transverse current setpoint frequency, respectively, whereby only one current component is examined depending on the injection direction. Switching is performed by the state-dependent switch 15.

[0071] If the measured quantities are highly noisy, a bandpass filter 11 can be used. A second-order bandpass filter is recommended, whereby the bandpass filter 11 should be applied to all measured quantities included in the calculation, so that the phase shift and amplitude attenuation occur equally for all quantities.

[0072] The Goertzel algorithm provides the DFT coefficients of the voltages and the injection current.

[0073] Shortly before the operating point changes (state 2 -> 0), the (differential) impedance of machine 1 is calculated from the DFT coefficients. The identified inductances are tracked to optimize the control system and stored in memory 14.

[0074] The controller parameters of the PI and P resonance controllers are automatically determined and adaptively tracked from measuring point to measuring point by the sequence control 3.

[0075] The test signal is interpreted automatically.

[0076] The dead time of inverter 8 is taken into account.

[0077] The differential inductances are calculated by dividing the DFT coefficients of current and voltage. Taking dead time into account, this allows for a highly accurate determination of the differential inductances and cross-couplings.

[0078] The flux linkages and the absolute fluxes can be calculated from the differential inductances.

[0079] The P-resonance controller can be temporarily deactivated to allow for a fast and stable adjustment of the DC components of the longitudinal and transverse flow.

Claims

[1] Method for determining current-dependent and / or rotation-angle-dependent parameters of an electrical machine (1) to be carried out using an inverter (8), comprising the steps: a) Setting a rotational angle position of a rotor of the electric machine (1) and subsequently locking the rotor, b) Forming a longitudinal flow setpoint by adding a longitudinal flow setpoint operating point (Id,0k) and a longitudinal current setpoint alternating signal (Id,HFk), where the longitudinal current setpoint alternating signal (Id,HFk) periodically changed with a longitudinal flow setpoint frequency, and / or forming a transverse flow setpoint by adding a transverse flow setpoint operating point (Iq,0k) and a cross-current setpoint alternating signal (Iq,HFk), where the cross-flow setpoint alternating signal (Iq,HFk) periodically changed with a cross-current setpoint frequency, c) Rules of a longitudinal flow (i d ) on the longitudinal flow setpoint and / or control of a transverse flow (i q ) to the cross-current setpoint, whereby control variables in the form of a longitudinal control voltage are used for regulation. (udk) and / or a transverse control voltage (uqk) generated and impressed into motor phases of the electric machine (1), and d) Measuring phase currents (i u,ist , i v,ist , i w,ist ) of the electric machine (1) and determining an emerging longitudinal current (i d, ist ) and / or an emerging crossflow (i q, ist ) from the measured phase currents (i u,ist , i v,ist , i w,ist ), characterized by the steps: e) Applying a discrete Fourier transform (DFT) algorithm or a Görtzel algorithm to the longitudinal control voltage (udk) and on the determined longitudinal flow (i d,ist ) to generate longitudinal control voltage coefficients and longitudinal current coefficients and / or to apply the DFT algorithm or the Görtzel algorithm to the transverse control voltage (uqk) and on the determined cross-flow (i q,ist ) for generating transverse control voltage coefficients and transverse current coefficients, and f) Calculating the parameters as a function of the longitudinal control voltage coefficients and the longitudinal current coefficients and / or the transverse control voltage coefficients and the transverse current coefficients, wherein the calculation of the parameters is carried out as a function of a dead time of the inverter (8) and / or as a function of signal propagation times of filters used to measure the phase currents. [2] Method according to claim 1, characterized by , that - in step b) the longitudinal flow setpoint operating point is changed and / or the transverse flow setpoint operating point is changed, with steps c) to f) being repeated when the longitudinal flow setpoint operating point is changed and / or when the transverse flow setpoint operating point is changed. [3] Method according to claim 1 or 2, characterized by , that - the control of longitudinal flow (i d ) on the longitudinal flow setpoint and / or the control of the transverse flow (i q ) is based on the cross-flow setpoint based on control parameters that depend on the longitudinal flow setpoint operating point and / or the cross-flow setpoint operating point. [4] Method according to claim 2 or 3, characterized by , that - the longitudinal flow setpoint operating point and / or the transverse flow setpoint operating point is / are continuously changed in such a way that a change in the longitudinal flow and / or the transverse flow does not exceed a threshold value. [5] Method according to any one of the preceding claims, characterized by , that - in step a) the rotation angle position is changed and then the rotor is blocked with the rotation angle position changed, whereby steps b) to f) are repeated with the rotation angle position changed. [6] Method according to any one of the preceding claims, characterized by , that - the current-dependent and / or rotation angle-dependent parameters are selected from the set of parameters: - differential longitudinal inductances, - differential transverse inductances, - Cross-saturation, - Resistors and - Impedances. [7] Method according to claim 6, characterized by the step: - Calculation of current-dependent and / or location-dependent flux linkages in longitudinal and transverse directions from the calculated differential longitudinal inductances and transverse inductances as well as the calculated cross-saturation, in the case of permanent magnet synchronous machines in particular taking into account the flux linkage in the longitudinal direction caused by their permanent magnets. [8] Method according to claim 7, characterized by the step: - Determining current-dependent and / or location-dependent absolute inductances in the longitudinal and transverse directions from the flux linkages in the longitudinal and transverse directions and from the associated longitudinal current (i d ) and crossflow (i q ). [9] Method according to any one of the preceding claims, characterized by , that - the longitudinal current obtained from the measurement of the phase currents (i d ) and crossflow (i q ) will be filtered. [10] Method according to any one of the preceding claims, characterized by , that - the electrical machine (1) is a synchronous machine, or a reluctance machine, or an induction machine. [11] Frequency converters comprising - a control unit designed to perform a method according to any one of claims 1 to 10.

Images