METHOD AND DEVICE FOR ESTIMATE THE ROTOR TEMPERATURE OF A ROTOR OF AN ELECTRIC MACHINE

Description

[0001] The invention relates to a method and a device for estimating the rotor temperature of a rotor of an electric machine.

[0002] Unlike internal combustion engines, electric drive systems are characterized by a significantly higher efficiency, which contributes substantially to the success of electromobility. Knowledge of the operating characteristics, particularly the rotor temperature, of an electric drive system allows for predictions regarding system aging. Furthermore, the system's operating limits, with respect to continuous power output and overload capacity, can be utilized more effectively. This results in crucial advantages for the traction characteristics of electrified vehicles. However, measuring the rotor temperature using a sensor is extremely difficult and expensive due to the high rotational speeds of electric motors in traction drives.

[0003] From M. Ganchev, C. Kral and T. Wolbank, Hardware and software implementation of sensorless rotor temperature estimation technique for Permanent Magnet Synchronous Motor, 2012 Electrical Systems for Aircraft, Railway and Ship Propulsion, 2012, pp. 1-6, doi: 10.1109 / ESARS.2012.6387420, a method for sensorless estimation of a rotor temperature is known. In this method, the estimation is performed while a q-current is regulated to zero.

[0004] From B. Weber, T. Brandt and A. Mertens, Compensation of switching dead-time effects in voltage-fed PWM inverters using FPGA-based current oversampling, 2016 IEEE Applied Power Electronics Conference and Exposition (APEC), Long Beach, CA, USA, 2016, pp. 3172-3179, doi: 10.1109 / APEC.2016.7468318, a method for dead time compensation is known.

[0005] From Bastian Weber, Positionsgeberlose Regelung von permanentmagneterexcitedn Synchronmaschinen bei kleinen Drehen mit überabprobender Stromerfassung, Hannover, Gottfried Wilhelm Leibniz Universität Hannover, Dissertation, 2018, 163 pp., https: / / doi.org / 10.15488 / 9140, a method for estimating a rotor position and / or a rotor speed is known.

[0006] DE 10 2015 005 555 A1 describes a method for determining the magnetic temperature of a permanent magnet electric machine, wherein at least one measured stator current of the machine, a stator voltage generated in a converter of the machine, and a measured rotational speed of the machine are supplied to a magnetic temperature monitoring unit, wherein the stator voltage is measured over a period of time using a model of the electric machine, wherein a permanent magnet flux is determined from the supplied data and taking into account a supply temperature of a cooling medium of the machine by means of the magnetic temperature monitoring unit, and a temperature difference is determined from a flux difference between the permanent magnet flux and a reference flux assigned to a reference magnet temperature and stored in at least one characteristic map in the model by means of the magnetic temperature monitoring unit, and the magnetic temperature is determined from this temperature difference.wherein frequency-dependent influences on the machine are determined offline to generate at least one characteristic map and are represented in it.

[0007] EP 2 894 784 B1 describes a magnet temperature estimation system for a synchronous electric motor with a permanent magnet, comprising a superposition unit configured to superimpose a voltage or current with a frequency different from the frequency of a fundamental shaft driving the synchronous electric motor onto at least one d-axis of the synchronous electric motor; a calculator configured to calculate an impedance of the synchronous electric motor from the superimposed voltage or current and a current or voltage obtained by the superposition; and a magnet temperature estimation unit configured to estimate a temperature of the permanent magnet based on the calculated impedance.

[0008] M. Ganchev, C. Kral, T. Wolbank, Sensorless rotor temperature estimation of permanent magnet synchronous motor under load conditions, Annual Conference of the IEEE Industrial Electronics Society, 2012, DOI:10.1109 / IECON.2012.6388895, describes a method for rotor temperature estimation.

[0009] The invention is based on the objective of creating a method and a device for estimating the rotor temperature of a rotor of an electric machine, with which the rotor temperature can be estimated more accurately.

[0010] The problem is solved according to the invention by a method with the features of claim 1 and a device with the features of claim 6. Advantageous embodiments of the invention are set forth in the dependent claims.

[0011] In particular, a method for estimating the rotor temperature of an electric machine rotor is provided, wherein phase currents of the electric machine are detected at high frequency, wherein, taking into account dead-time compensation, the associated phase voltages are determined, wherein a rotor position of the electric machine is detected or estimated, wherein, starting from the high-frequency detected phase currents, the determined phase voltages and the detected or estimated rotor position, a voltage component of the stator voltages caused by the flux linkage of the rotor is determined, wherein the flux linkage of the rotor is determined starting from the determined voltage component, and wherein the rotor temperature of the electric machine is estimated and provided based on a known dependence of the rotor temperature on the flux linkage of the rotor.

[0012] Furthermore, a device for estimating the rotor temperature of an electric machine rotor is provided, comprising a control device, wherein the control device is configured to receive high-frequency detected phase currents, to determine associated phase voltages taking into account dead-time compensation, to receive a detected or estimated rotor position of the electric machine, to determine a voltage component of the stator voltages caused by the flux linkage of the rotor based on the high-frequency detected phase currents, the determined phase voltages and the detected or estimated rotor position, to determine the flux linkage of the rotor based on the determined voltage component, and to estimate and provide the rotor temperature of the electric machine based on a known dependence of the rotor temperature on the flux linkage of the rotor.

[0013] The method and device make it possible to estimate the rotor temperature even with a q-current greater than zero, i.e., during normal operation of the electric machine. This allows the rotor temperature to be estimated even when torque is being demanded from the electric machine. Therefore, the torque-generating current control does not need to be interrupted for the purpose of determining the rotor temperature. Since no interruption is required, a drive system implemented using the electric machine offers the advantage of greater torque dynamics compared to systems that require an interruption of the torque-generating current (q-current) to determine the rotor temperature.

[0014] The advantage is achieved primarily through high-frequency acquisition of the phase currents and dead-time compensation. Specifically, the phase currents of the electric machine are continuously acquired at a high frequency. This high-frequency acquisition means that the phase currents are oversampled. In particular, this means that the sampling frequency is significantly higher than the modulation frequency of a pulse inverter used to power the electric machine, especially by a factor of 10, 100, or 1000. Taking dead-time compensation into account, the corresponding phase voltages are determined. This provides current values ​​for both the phase currents and their respective phase voltages. The rotor position of the electric machine is also acquired or estimated.Based on the high-frequency measured phase currents, the corresponding phase voltages, and the measured or estimated rotor position, a voltage component of the stator voltages caused by the rotor flux linkage is determined. The rotor flux linkage is then determined from this voltage component, and the rotor temperature of the electric machine is estimated and provided based on a known relationship between the rotor temperature and the rotor flux linkage. This provision can, for example, involve generating and outputting an analog or digital rotor temperature signal.

[0015] The phase currents are detected and provided, in particular by means of suitable sensors, especially current sensors on the individual phases. The high-resolution detected phase currents can be used, in particular, to generate phase current values ​​equivalent to center-synchronous sampling. This high-resolution detection allows for improved noise filtering in the equivalent values, thus improving the signal-to-noise ratio.

[0016] The string voltages can be determined, in particular, based on the measured string currents and a modulation control of the pulse inverter. Dead-time compensation is provided for, which is described in more detail below.

[0017] A rotor position is, in particular, the rotor angle. A rotor position can be detected and provided, for example, by means of appropriate sensors, such as a rotary encoder.

[0018] The control unit can be implemented individually or as a combination of hardware and software, for example, as program code running on a microcontroller or microprocessor. However, it can also be designed with components implemented individually or as an application-specific integrated circuit (ASIC) or field-programmable gate array (FPGA).

[0019] It is specifically intended that the high-frequency acquired, especially oversampled, phase currents, and optionally other quantities, are evaluated using a field-programmable gate array (FPGA). A field-programmable gate array enables high-frequency signal acquisition and processing. "High frequency" here means, in particular, that the corresponding frequency range lies above the modulation frequency of a pulse-width modulation of the pulse inverter. Specifically, "high frequency" refers to a frequency that is 10 times, preferably 100 times, and most preferably 1000 times higher than the modulation frequency. For example, if the modulation frequency is 20 kHz, then the signal acquisition and processing frequency of the field-programmable gate array is 1 MHz.Through the parallel signal processing of a field-programmable gate array, it can advantageously process high-frequency (especially ~MHz) sampled quantities. High-frequency sampling reduces sampling and aliasing effects as noise in the signal. Any noise components from the sensor can be digitally filtered in the field-programmable gate array, thus providing a very high signal-to-noise ratio for further digital processing. In comparison, a microcontroller (e.g., a DSP) is significantly less suitable for processing high-frequency signals, as the acquired values ​​of the DSP must be transported from the acquiring internal analog-to-digital converter (ADC) to the processor core and processed, which leads to delays in evaluation due to signal propagation time and preprocessing.In a field-programmable gate array, the acquired values ​​from the ADC are immediately available to the logic units, i.e., in real time. Field-programmable gate arrays also allow for the evaluation of cost-effective ADCs capable of generating a high-frequency digital data stream (bitstream), such as delta-sigma converters, without further processing. The field-programmable gate array is typically a component of a system-on-a-chip (SoC).

[0020] The field-programmable gate array can, in particular, be a field-programmable gate array that is partially reconfigurable at runtime. This makes the field-programmable gate array particularly updateable. Reconfiguration is achieved, in particular, by means of partial reconfiguration, where only a portion of the gate array is reconfigured or reprogrammed at runtime, while other parts of the gate array retain their respective configuration or programming. Reconfiguration is carried out, in particular, by means of a configuration controller, which performs and monitors the reconfiguration. The configuration controller can be provided by means of a DSP or the control unit of the device.To reconfigure at least one field-programmable gate array, methods for automatic code generation can be used, such as the "Xilinx System Generator for DSP," which can automate both source code modeling and functional analysis. The "System Generator for DSP" enables, for example, the modeling of FPGA functions in a MATLAB Simulink environment, allowing their properties to be tested against other Simulink function blocks. Automatic code generation ensures that the field-programmable gate array behaves identically to the Simulink model. A field-programmable gate array configuration includes, in particular, instructions that program the array so that the programmed (or...)A specific data processing functionality can be provided to a configured or reconfigured part of the field-programmable gate array. In the present case, such functionality is, in particular, the execution of measures of the procedure described in this disclosure.

[0021] It is intended that dead-time effects occurring in the pulse inverter due to switching delay times are compensated for by means of dead-time compensation. This is achieved in particular by means of the control unit, specifically by means of at least one field-programmable gate array of the control unit. The dead-time compensation attempts to compensate for a dead-time effect. This dead-time effect occurs because, due to the switching times of the power semiconductors in the pulse inverter, there is a misrepresentation between an estimated current voltage and a real voltage. To compensate for the dead-time effect, the switching edges of the power semiconductors are adjusted, that is, shifted with respect to a switching point. This allows the actual current phase voltages to be precisely determined, and in particular estimated, taking into account the high-frequency phase currents. The dead-time compensation is implemented, for example, by means of the [method / component] described in B.Weber, T. Brandt and A. Mertens, Compensation of switching dead-time effects in voltage-fed PWM inverters using FPGA-based current oversampling, 2016 IEEE Applied Power Electronics Conference and Exposition (APEC), Long Beach, CA, USA, 2016, pp. 3172-3179, doi: 10.1109 / APEC.2016.7468318, carried out the procedure described.

[0022] The procedure for determining the rotor temperature can be illustrated using the stator voltage differential equation of a permanent magnet synchronous machine. For asynchronous machines, the procedure is fundamentally analogous.

[0023] The permanent magnet synchronous machine can be modeled in stator-fixed αβ coordinates by the following stator voltage differential equations: u α u β = R ⋅ i α i β + d dt Σ L 1 0 0 1 + Δ L cos 2 γ el sin 2 γ el sin 2 γ el − cos 2 γ el ⋅ i α i β + d dt ψ PM cos 2 γ el sin 2 γ el

[0024] These are: Σ L = L d + L q 2 und Δ L = L d − L q 2 , L d the inductance of the electric machine with respect to the d-axis of the rotor (direct axis), L qthe inductance of the electric machine with respect to the q-axis of the rotor (quadrature axis), R the phase resistance of the electric machine, ψ PM the flux linkage of the rotor due to the permanent magnets, γ el the electrical rotor position (i.e. the electrical rotor position angle), u α , u β the stator voltages in stator-fixed αβ coordinates and i α , i β the stator currents in stator-fixed αβ coordinates.

[0025] The phase voltages and phase currents can be converted into the stator voltages and stator currents in αβ coordinates.

[0026] The inductance values L d and L q These values ​​are almost entirely independent of temperature and can be measured, for example before applying the method, or are known for a series of a particular type of electric machine. The phase resistance R and the flux linkage are particularly temperature-dependent. ψ PM .R typically increases with temperature and ψ PM decreases with temperature.

[0027] Basically, the method described in this disclosure is about reducing the resistive-inductive voltage component of the stator voltages: R ⋅ i α i β + d dt Σ L 1 0 0 1 + Δ L cos 2 γ el sin 2 γ el sin 2 γ el − cos 2 γ el ⋅ i α i β to separate from a voltage component of the rotor caused by flux linkage: d dt ψ PM cos 2 γ el sin 2 γ el

[0028] This is done based on the high-frequency detected string currents, the determined string voltages, and the detected or estimated rotor position.

[0029] The method and the device can be used, for example, in a motor vehicle, particularly an electric or hybrid vehicle. In principle, however, the method and the device can also be used in other areas, especially in other land, rail, water, air, or space vehicles, such as drones or air taxis.

[0030] In an unclaimed embodiment, the stator temperature is detected by means of at least one temperature sensor. Based on the detected stator temperature, the phase resistance of the electrical machine is estimated. To determine the flux linkage, the flux linkage is estimated by an observer, taking the estimated phase resistance into account. The phase resistance of the electrical machine is estimated, for example, using a characteristic curve in which stator temperature values ​​are correlated with phase resistance values. Such a characteristic curve can be determined empirically and / or by simulation. If the phase resistance is known, the flux linkage can be estimated by an observer using known inductances of the electrical machine, which may have been measured beforehand.In particular, the observer can be used to estimate the time course of the flux linkage as a state variable. An observer is specifically a control engineering observer (state observer) that estimates the state of the electric machine based on a model of the machine. The estimated state of the electric machine is compared with the actual state of the electric machine, and if there are deviations, at least one parameter of the model is adjusted until the estimated state and the actual state match again. In this case, the adjusted parameter is specifically the flux linkage. The rotor temperature of the electric machine is then estimated and provided based on the known dependence of the rotor temperature on the rotor flux linkage.

[0031] It may be possible to determine and provide an average rotor temperature from the estimated rotor temperature. This allows fluctuations in rotor temperature over time to be compensated for.

[0032] It is provided that a stator temperature is detected by means of at least one temperature sensor, and based on the detected stator temperature, the phase resistance of the electric machine is estimated. Zero-voltage states are identified based on the current gradients of the detected phase currents. Based on the detected current gradients, known values ​​for the phase voltages during the zero-voltage states, and the estimated phase resistance, the voltage component of the flux linkage is determined. During the zero-voltage states, the electric machine is short-circuited. Forward voltage components of power semiconductors of a pulse inverter supplying the electric machine are nearly zero and can therefore be neglected, so that the stator voltages u α , u βBoth are zero. Based on the current gradients, which can be detected particularly due to the high-frequency or oversampled acquisition of the phase currents with a high signal-to-noise ratio, it is d dt Σ L 1 0 0 1 + Δ L cos 2 γ el sin 2 γ el sin 2 γ el − cos 2 γ el ⋅ i α i β also known. Since the stator temperature is measured, it can R ⋅ i α i β The stress component of the flux linkage can therefore be determined from the known values. The flux linkage is then determined from this stress component, and the rotor temperature is estimated from this, as already described.

[0033] In an unclaimed embodiment, a high-frequency test signal with respect to the phase voltages is applied by modulating a pulse inverter supplying the electric machine. To determine the flux linkage using an observer, a flux linkage voltage component in the system response of the electric machine to the applied high-frequency test signal is separated from a resistive-inductive voltage component in the system response. The high-frequency test signal is applied to the electric machine by the voltage generated by the pulse inverter. However, this high-frequency test signal is selected to be so short (in particular, approximately 1 to 10 pulse periods of the pulse inverter) that the back EMF of the electric machine can be approximated as nearly constant. The resistive-inductive voltage component reacts with a system response at the same frequency. R ⋅ i α i β + d dt Σ L 1 0 0 1 + Δ L cos 2 γ el sin 2 γ el sin 2 γ el − cos 2 γ el ⋅ i α i β whereas the voltage component of the rotor is a constant quantity: d dt ψ PM cos 2 γ el sin 2 γ el

[0034] This will be explained in more detail below:

[0035] The voltage component of the rotor is determined by its rotational speed, which is subject to mechanical inertia. This voltage component can therefore be considered a DC quantity over short periods, particularly approximately 1 to 10 pulse duration periods of the pulse inverter. If a high-frequency test signal is now applied to the phase voltages of the electric machine by modulating the pulse inverter, the aforementioned resistive-inductive voltage component exhibits a DC voltage response. With an observer, for example by evaluating entire periods of a carrier frequency of the test signal, the voltage component of the flux linkage in the system response of the electric machine to the applied high-frequency test signal can thus be separated from the resistive-inductive voltage component in the system response. An observer is, in particular, a control engineer (i.e., a system controller).A state observer, based on a model of the electric machine, estimates the state of the electric machine. The estimated state is compared with the actual state of the electric machine, and if there are deviations, at least one parameter of the model is adjusted until the estimated state and the actual state match again. In this case, the adjusted parameter is, in particular, the flux linkage. ψ PM or a temporal progression of the river chain ψ PM The rotor temperature of the electric machine is then estimated and provided based on the known dependence of the rotor temperature on the flux linkage of the rotor.

[0036] In an unclaimed further developmental embodiment, it is provided that the phase currents and phase current profiles are acquired at high frequency during periods of active voltage states or reconstructed from the phase current profile during periods of zero voltage states. Evaluating the phase currents during active voltage states is particularly advantageous at operating points with high modulation levels of the pulse inverter, for example, at high rotational speeds.

[0037] In one embodiment, the control or regulation of the electric machine is adapted based on the estimated rotor temperature. This allows the control and / or regulation to be directly adjusted based on a current value for the rotor temperature, ensuring that the electric machine is always controlled and / or regulated based on current parameter values.

[0038] In one embodiment, the control system allows for adjustments to the maximum continuous power output and / or the dynamic overload capacity of the electric machine. Specifically, related operating parameters are modified. This enables the electric machine to operate at its maximum possible or achievable maximum continuous power output and / or maximum possible dynamic overload capacity, depending on a current or average value. In particular, this improves both performance and service life at any given time. In other words, it prevents damaging overloads at all times, thus extending the service life.

[0039] In one embodiment, the rotor position is estimated as part of a sensorless control system for the electric machine. A method for estimating a rotor position and / or rotor speed is known, for example, from Bastian Weber, Positionsgeberlose Regelung von permanentmagneterrangten Synchronmaschinen bei kleinen Drehen mit überabprobender Stromerfassung [Sensorless Control of Permanent Magnet Excited Synchronous Machines at Low Speeds with Oversampling Current Detection], Hannover, Gottfried Wilhelm Leibniz University Hannover, Dissertation, 2018, 163 pp., https: / / doi.org / 10.15488 / 9140.

[0040] In a further developed embodiment, the rotor position estimation of the sensorless control is deactivated when a torque demand on the electric machine falls below a threshold value and / or is zero. After deactivation, the rotor position is extrapolated from previously estimated rotor positions. The rotor temperature is estimated while the rotor position estimation is deactivated, and the rotor position estimation is reactivated after the rotor temperature has been estimated. This avoids any potentially disruptive influence of the rotor position estimation of the sensorless control, which might occur, for example, because not all necessary parameters can be determined simultaneously. In other words, intermittent operation is performed when torque demands are very low or negligible, during which the measures for determining the rotor temperature are carried out.If the torque demand on the electric machine is very low or zero for a short period of approximately four PWM cycles, the rotor position estimation of the sensorless control system is temporarily deactivated. Due to the mechanical inertia of the electric machine's rotor, it is assumed that the rotational speed does not change during this period, allowing the current rotor position to be extrapolated from previously known values. With the sensorless control system deactivated, the rotor temperature is estimated. The sensorless control system is then reactivated and resynchronizes with the electric machine.

[0041] Further features for the design of the device result from the description of embodiments of the method. The advantages of the device are the same in each case as in the embodiments of the method.

[0042] In particular, it may be provided that the control device is at least partially designed as a field-programmable gate array and / or includes at least one field-programmable gate array.

[0043] Further development may stipulate that the control unit includes at least one system-on-chip.

[0044] The invention is explained in more detail below with reference to preferred embodiments and the figures. These show: Fig. 1 a schematic representation of an embodiment of the device for estimating the rotor temperature of a rotor of an electric machine; Fig. 2 a schematic representation to illustrate high-frequency detection of phase currents; Fig. 3 a schematic flowchart of an embodiment of the method for estimating the rotor temperature of a rotor of an electric machine.

[0045] In Fig. 1Figure 1 shows a schematic representation of an embodiment of the device 1 for determining the rotor temperature 40 of a rotor of an electric machine 20. The device 1 comprises a control unit 2. In the embodiment shown, the control unit 2 comprises a digital signal processor 3 and a field-programmable gate array 4. The field-programmable gate array 4 allows high-frequency signal processing and evaluation. The digital signal processor 3 and the field-programmable gate array 4 are, in particular, designed as a system-on-chip (SoC). The device 1 performs the method described in this disclosure, which is explained below with reference to the device 1.

[0046] The field-programmable gate array 4 comprises several modules 4-x that perform different functions, as explained below.

[0047] Also shown is a pulse inverter 10, which in this example is configured as a two-point inverter, as well as a sensor system 12 for detecting phase currents ix, comprising current sensors 12-1, 12-2, 12-3 and an analog-to-digital converter 12-4. The sensor system 12 can also be part of the device 1. Furthermore, an electric machine 20 is shown, which is supplied by the pulse inverter 10 via three phases 11-1, 11-2, 11-3. For supply, the pulse inverter 10 generates a rotating magnetic field in a manner known per se and is controlled for this purpose by a module 4-1. In particular, the module 4-1 provides modulation control.

[0048] The phase currents ix of the electric machine 20 are detected at high frequency by means of the sensor 12, received by the control unit 2, and processed by the module 4-2. In particular, the phase currents ix are oversampled at a sampling frequency that is 10, 100, or even 1000 times higher than the modulation frequency of the pulse inverter 10. Taking into account dead-time compensation, the corresponding phase voltages are also determined by means of the module 4-2.

[0049] A rotor position 14, i.e., an electrical rotor angle, of the electric machine 20 is detected by means of a rotary encoder 21, received by the control unit 2 and fed to the module 4-2, or, in the context of an encoderless control system, estimated from the detected phase currents ix. The estimation is performed, for example, using the module 4-2.

[0050] Starting from the high-frequency detected phase currents ix, the determined phase voltages, and the detected or estimated rotor position 14, module 4-2 determines a voltage component of the stator voltages of the electric machine 20 caused by the rotor's flux linkage 32. A value for the rotor's flux linkage 32 is determined by module 4-2 based on this determined voltage component. The rotor temperature 40 of the electric machine 20 is estimated based on a known dependence of the rotor temperature 40 on the rotor's flux linkage 32, for example, by module 4-2 or 4-3, and is provided, for example, by module 4-3. The estimated rotor temperature 40 is provided, for example, as an analog or digital signal, for example, in the form of a digital data packet, and in particular, output.

[0051] It may be possible to determine and provide an average rotor temperature 41 from several values ​​of the estimated rotor temperature 40. The averaging is carried out, for example, using module 4-3.

[0052] Module 4-3 provides in particular a control of the electrical machine 20, whereby the control loop is realized together with modules 4-1 and 4-2.

[0053] In an unclaimed embodiment, it may be provided that a stator temperature 19 is detected by means of a temperature sensor 22, wherein, starting from the detected stator temperature 19, a phase resistance 31 of the electric machine 20 is estimated, wherein, to determine the flux linkage 32, the flux linkage 32 is estimated taking into account the estimated phase resistance 31 by means of an observer 5. The observer 5 is provided by means of module 4-3.

[0054] It is provided that the stator temperature 19 is detected by means of the temperature sensor 22, wherein, starting from the detected stator temperature 19, a phase resistance 31 of the electrical machine 20 is estimated, wherein, starting from current gradients of the detected phase currents ix, zero voltage states are identified, wherein, starting from the detected current gradients, known values ​​for the phase voltages during the zero voltage states and the estimated phase resistance 31, the voltage component of the flux linkage 32 is determined.

[0055] Alternatively, in an unclaimed embodiment, it can be provided that a high-frequency test signal 30 with respect to the phase voltages is impressed by modulation of the pulse inverter 10 supplying the electric machine 20, wherein, to determine the flux linkage 32, a voltage component of the flux linkage 32 in a system response of the electric machine 20 to the impressed high-frequency test signal 30 is separated from a resistive-inductive voltage component in the system response by means of an observer 5. For this purpose, the pulse inverter 10 is controlled by the module 4-1 such that it generates and impresses a high-frequency test signal 30 with respect to the phase voltages by modulation. For this purpose, the pulse inverter 10 is controlled by means of the module 4-1 in such a way that a high-frequency oscillation, especially in the range of 3 to 5 kHz or above, for example with an amplitude of 10 V, is impressed into the d-axis.This stimulates all three strands 11 times.

[0056] In the unclaimed embodiment, it may be further provided that the phase currents ix and phase current profiles over time are recorded at high frequency during a period of active voltage states or are reconstructed from a profile of the phase currents ix during the zero voltage states.

[0057] It may be provided that, based on the estimated rotor temperature 40, the control or regulation of the electric machine 20 is adjusted. For this purpose, the associated parameters of the control or regulation are changed. These parameters can, for example, be adjusted based on a characteristic curve 45 stored in the control unit 2, in which values ​​for the (averaged) rotor temperature 40 (41) are linked to the respective values ​​for the parameters of the control or regulation.

[0058] It may be provided, in particular, that the maximum continuous power 46 and / or the dynamic overload capacity 47 of the electric machine 20 are modified for adjustment in the control or regulation system. This is also done, for example, based on motor characteristics (not shown) of the electric machine 20, which link rotor temperature values ​​40 with values ​​for the maximum continuous power 46 and / or the dynamic overload capacity 47. The modification is carried out, for example, in module 3-1 of the digital signal processor 3. The modified values ​​for the maximum continuous power 46 and / or the dynamic overload capacity 47 are supplied to module 4-3, which controls or regulates the electric machine 20. Module 4-3 adjusts the control or regulation accordingly. In particular, an operating parameter range of the electric machine 20 is adjusted according to the modified limits.

[0059] When estimating the rotor position 14 within the framework of sensorless control, it may be provided that a rotor position estimation of the sensorless control is deactivated if a torque request to the electric machine 20 falls below a threshold value and / or is zero, wherein a rotor position 14 is extrapolated from previously estimated rotor positions 14 after deactivation, wherein the rotor temperature 40 is determined while the rotor position estimation is deactivated, and wherein the rotor position estimation is reactivated after the rotor temperature 40 has been determined.

[0060] In Fig. 2Figure 1 shows a schematic representation illustrating the high-frequency acquisition of phase currents ix. Exemplary time profiles of a phase current ix, a phase voltage ux, and a carrier signal of a pulse-width modulation of a pulse inverter are shown over time t, each over a complete modulation period T of the pulse inverter. Both active voltage states 17 and zero-voltage states 18 are shown.

[0061] In the commonly used center-synchronous sampling, two measured values ​​15 of the phase current ix are acquired per period, namely at the beginning and in the middle of the modulation period T. In contrast, with high-frequency acquisition of the phase current ix, significantly more measured values ​​16 are acquired per modulation period T (for clarity, only a few are shown with their own reference symbol). In particular, oversampling takes place, meaning that a sampling frequency is, in particular, 10, 100, or even 1000 times higher than the modulation frequency of the pulse inverter. Therefore, a multiple of measured values ​​16 for the phase current ix is ​​available per modulation period T, which has a positive effect on the signal-to-noise ratio and also enables averaging and / or filtering. Furthermore, real-time measured values ​​16 are always available throughout the entire modulation period T.In particular, current gradients can also be determined more accurately.

[0062] By recording the string currents ix at high frequency, it is particularly possible to perform dead time compensation, so that the real current string voltages ux, which belong to the respective string currents, can be determined.

[0063] Furthermore, high-frequency detection makes it possible in particular to detect a phase current ix and a phase current profile over time at high frequency during a period of active voltage states 17.

[0064] Furthermore, high-frequency acquisition makes it possible, in particular, to alternatively acquire a phase current ix during the zero-voltage states 18 and to reconstruct the phase current ix during the active voltage states 17 from the phase current ix acquired during the zero-voltage states 18 by estimation. In the simplest case, the values ​​during the active voltage states 17 are interpolated from the values ​​before and after the active voltage states 17.

[0065] In one embodiment of the method and the device, it is provided that the phase currents ix and phase current profiles over time are recorded at high frequency during a duration of the active voltage states 17 or are reconstructed from a profile of the phase currents during the zero voltage states 18.

[0066] In Fig. 3A schematic flowchart of an embodiment of the method for estimating the rotor temperature of a rotor of an electric machine is shown.

[0067] In measure 100a, the phase currents of the electric machine are recorded at a high frequency. This is done in particular by oversampling with a sampling frequency that is 10, 100 or 1000 times greater than the modulation frequency of a pulse inverter that feeds the electric machine.

[0068] In measure 100b, the corresponding string voltages are determined, taking into account dead-time compensation. Specifically, a corresponding string voltage is determined at each time a value for a string current is recorded. This is done primarily based on the recorded string currents and a modulation of the pulse inverter.

[0069] In measure 100c, the rotor position of the electric machine is detected or estimated. This can be done using a dedicated sensor system, for example a rotary encoder, or as part of a encoderless control system where the rotor position is estimated.

[0070] Measures 100a to 100c are repeated continuously and, in particular, carried out simultaneously, so that current values ​​for the string currents, the string voltages and the rotor position are always available.

[0071] In measure 101, based on the high-frequency recorded string currents, the determined string voltages and the recorded or estimated rotor position, a voltage component of the stator voltages caused by the flux linkage of the rotor is determined.

[0072] In measure 102, the flux linkage of the rotor is determined starting from the voltage component determined in measure 101.

[0073] In measure 103, the rotor temperature of the electric machine is estimated and provided based on a known dependence of the rotor temperature on the rotor flux linkage. The estimated rotor temperature can be provided, for example, as an analog or digital rotor temperature signal, and in particular, output.

[0074] In an unclaimed embodiment, it may be provided that in measure 100d a stator temperature is detected by means of at least one temperature sensor. In measure 101, it is then provided that, starting from the detected stator temperature, a phase resistance of the electrical machine is estimated, wherein in measure 102, to determine the flux linkage, the flux linkage is estimated by means of an observer, taking into account the estimated phase resistance.

[0075] It is provided that in measure 100d, a stator temperature is measured using at least one temperature sensor. In measure 101, the phase resistance of the electrical machine is then estimated based on the measured stator temperature. In measure 102, zero-voltage states are identified based on the current gradients of the measured phase currents. Based on the measured current gradients, known values ​​for the phase voltages during the zero-voltage states, and the estimated phase resistance, the voltage component of the flux linkage is determined.

[0076] Alternatively, in an unclaimed embodiment, it may also be provided that in measure 101 a high-frequency test signal with respect to the phase voltages is applied by modulating a pulse inverter supplying the electric machine. In measure 102, to determine the flux linkage, a voltage component of the flux linkage in a system response of the electric machine to the applied high-frequency test signal is then separated from a resistive-inductive voltage component in the system response by means of an observer.

[0077] The unclaimed embodiment may be further developed by recording the phase currents and phase current profiles at high frequency during periods of active voltage states or by reconstructing them from a profile of the phase currents during periods of zero voltage states.

[0078] Measure 104 may provide for the adjustment of a control or regulation of the electric machine based on the estimated rotor temperature.

[0079] Further training measures, specifically measure 104, may include changing the maximum continuous power output and / or the dynamic overload capacity of the electrical machine in order to adapt it in the control or regulation system.

[0080] It may be provided that a rotor position estimation of the sensorless control is deactivated in a measure 99 if a torque request to the electric machine falls below a threshold value and / or is zero, wherein a rotor position after deactivation in measure 100c is extrapolated from previously estimated rotor positions, wherein the rotor temperature is estimated by means of measures 100a to 103 while the rotor position estimation is deactivated, and wherein the rotor position estimation is reactivated after estimating the rotor temperature in a measure 105. Reference symbol list

[0081] 1 device 18 Zero voltage state 2 Control unit 19 Stator temperature 3 Digital signal processor 20 electric machine 4 field-programmable gate array 21 Rotary encoder 22 temperature sensor 4-x module 30 high-frequency test signal 5 observer 31 String resistance 10 Pulse inverter 32 River chain 11-x strand 40 Rotor temperature 12 Sensors 41 average rotor temperature 12-1 Current sensor 45 characteristic curve 12-1 Current sensor 46 Maximum continuous power 12-1 Current sensor 47 dynamic overload capacity 12-4 Digital-to-analog converter 99-105 measure 14 Rotor position ix String current 15 Measurement ux String voltage 16 Measurement t Time 17 active voltage state T Modulation period

Claims

1. Method for estimating a rotor temperature (40) of a rotor of an electric machine (20), wherein phase currents (i-x) of the electric machine (20) are detected at high frequency, i.e. on an oversampled basis, wherein associated phase voltages (u-x) are determined taking dead-time compensation into account, wherein a rotor position (14) of the electric machine (20) is detected or estimated, wherein a voltage component of stator voltages caused by the flux linkage (32) of the rotor is determined on the basis of the phase currents (i-x) detected at high frequency, the determined phase voltages (u-x) and the detected or estimated rotor position (14), wherein the flux linkage (32) of the rotor is determined on the basis of the determined voltage component, and wherein the rotor temperature (40) of the electric machine (20) is estimated and provided on the basis of a known dependence of the rotor temperature (40) on the flux linkage (32) of the rotor, wherein a stator temperature (19) is detected by means of at least one temperature sensor (22), wherein a phase resistance (31) of the electric machine (20) is estimated on the basis of the detected stator temperature (19), wherein zero-voltage states (18) are identified on the basis of current increases of the detected phase currents (i-x), wherein the voltage component of the flux linkage (32) is determined on the basis of the detected current increases, known values for the phase voltages (u-x) during the zero-voltage states and the estimated phase resistance (31).

2. Method according to any of the preceding claims, characterized in that open-loop or closed-loop control of the electric machine (20) is adjusted on the basis of the estimated rotor temperature (40).

3. Method according to claim 2, characterized in that a maximum continuous power and / or a dynamic overload capacity of the electric machine (20) are changed for adjustment in the open-loop or closed-loop control.

4. Method according to any of the preceding claims, characterized in that the rotor position (14) is estimated within the context of sensorless closed-loop control of the electric machine (20).

5. Method according to claim 4, characterized in that rotor position estimation of the sensorless closed-loop control is deactivated when a torque request to the electric machine (20) does not meet a threshold value and / or is zero, a rotor position (14) being extrapolated from previously estimated rotor positions (14) after deactivation, the rotor temperature (40) being estimated while the rotor position estimation is deactivated, and the rotor position estimation being reactivated after the rotor temperature (40) has been estimated.

6. Device (1) for estimating a rotor temperature (40) of a rotor of an electric machine (20), comprising: at least one temperature sensor (22), designed to detect a stator temperature (19), and a control apparatus (2), wherein the control apparatus (2) is designed to receive phase currents (i-x) detected at high frequency, i.e. on an oversampled basis, to determine associated phase voltages (u-x) taking dead-time compensation into account, to receive a detected or estimated rotor position (14) of the electric machine (20), to determine a voltage component of stator voltages caused by the flux linkage (32) of the rotor on the basis of the phase currents (i-x) detected at high frequency, the determined phase voltages (u-x) and the detected or estimated rotor position (14), to determine the flux linkage (32) of the rotor on the basis of the determined voltage component, and to estimate and provide the rotor temperature (40) of the electric machine (20) on the basis of a known dependence of the rotor temperature (40) on the flux linkage (32) of the rotor, and to estimate a phase resistance (31) of the electric machine (20) on the basis of the detected stator temperature (19), to identify zero-voltage states (18) on the basis of current increases in the detected phase currents (i-x), and to determine the voltage component of the flux linkage (32) on the basis of the detected current increases, known values for the phase voltages (u-x) during the zero-voltage states and the estimated phase resistance (31).