DE102015005555A1 - Method for determining a magnet temperature of a permanently excited electrical machine - Google Patents

Abstract

The invention relates to a method for determining a magnet temperature (θ ^ PM) of a permanently excited electric machine, wherein a magnet temperature observing unit (1) at least one measured stator current (idq [k]) of the machine, a stator voltage formed in a converter of the machine (udq k]) and a measured speed of the machine are supplied, wherein a detection of the stator voltage (udq [k]) over a runtime on the basis of a model (2) of the electric machine, wherein by means of the magnetic temperature observer unit (1) from the supplied data and Taking into account a flow temperature of a cooling medium of the machine, a permanent magnet flux (ψ ^ dq [k]) is determined and - by the magnetic temperature observer unit (1) from a flow difference (Δψ ^ dq [k]) between the permanent magnet flux (ψ ^ dq [k]) and a reference associated with a reference magnet temperature (θ PM, 0) and stored in the model (2) in at least one map) flow temperature difference (Δθ ^ PM) and from this the magnet temperature (θ ^ PM) is determined, where - to generate the at least one map) frequency-dependent influences on the machine are determined offline and in the at least one map) are mapped.

Description

The invention relates to a method for determining a magnet temperature of a permanently excited electrical machine.

For an electric drive of vehicles mainly permanent magnetically excited electrical machines are used due to a high power density and a variable design. Due to external heat inputs, for example due to the stator of the electric machine or other assemblies, which are located in an environment of the machine, and due to iron losses caused self-heating occurs in the operation of such machines to a significant heating of the rotor. In the rotor are permanent magnets, which have a significant temperature dependence in terms of their material properties.

• – Im zulässigen Temperaturbereich der Permanentmagnete findet eine reversible Entmagnetisierung statt, welche direkten Einfluss auf das Drehmoment der Maschine hat. Mithilfe einer entsprechenden Temperaturinformation kann eine Parametervariation in einer Regelung nachgeführt und eine Drehmomentgenauigkeit erhöht werden. Dies ist unter anderem vorteilhaft für eine aktive Schwingungsdämpfung im mechanischen Antriebsstrang.
• – Außerhalb des zulässigen Temperaturbereichs findet eine irreversible Entmagnetisierung statt, welche zu einer dauerhaften Leistungsminderung der Maschine führt. In diesem Fall ist ein Austausch der Maschine erforderlich. Durch die Kenntnis der Magnettemperatur zur Laufzeit der Maschine kann die Betriebssicherheit erhöht werden.
• – Weiterhin wird eine höhere thermische Ausnutzung der Maschine ermöglicht. Um die temperaturbedingte irreversible Entmagnetisierung sicher zu verhindern, werden typischerweise so genannte statische Derating-Kennlinien eingesetzt, welche mit großen Sicherheitsaufschlägen behaftet sind. Durch die Kenntnis der Magnettemperatur zur Laufzeit kann eine Überlastfähigkeit der Maschine besser ausgenutzt werden, da ein dynamisches Derating bei geschlossenem Temperaturregelkreis möglich wird.

In particular, for the following reasons, it is necessary to know a magnet temperature during operation of the machine:

• - In the permissible temperature range of the permanent magnets, a reversible demagnetization takes place, which has a direct influence on the torque of the machine. By means of a corresponding temperature information, a parameter variation can be tracked in a control and a torque accuracy can be increased. This is advantageous, inter alia, for active vibration damping in the mechanical drive train.
• - Outside the permissible temperature range irreversible demagnetization takes place, which leads to a permanent reduction in performance of the machine. In this case, an exchange of the machine is required. By knowing the magnet temperature at runtime of the machine, the reliability can be increased.
• - Furthermore, a higher thermal utilization of the machine is possible. In order to reliably prevent the temperature-induced irreversible demagnetization, so-called static derating characteristics are typically used, which are associated with large safety impacts. By knowing the magnet temperature at runtime, an overload capacity of the machine can be better utilized, as a dynamic derating is possible with closed temperature control loop.

Since a direct measurement of the magnet temperature in the rotating rotor is very expensive, the magnet temperature is generally determined indirectly. For this purpose, three different methods are generally known from the prior art.

In a first method, a thermal modeling takes place in which a thermal behavior of the machine is mapped by means of an abstract model based on lumped parameter thermal network (LPTN for short). The parameters of the model are identified on the basis of experimental test data or analytically determined on the basis of geometric considerations. Magnetic temperature determination takes place as an open estimate using the previously established abstract model (LPTN).

In a second method, also referred to as signal injection, it is assumed that a variation of the permanent magnet flux as a function of the temperature has an effect on an electrical-magnetic machine behavior. By applying voltage signals, it is possible to conclude by means of a current response to the magnet temperature. In this case, either an iron saturation characteristic or a machine impedance at high-frequency excitation is evaluated.

A third method, also referred to as an accurate flux observer, is based on the assumption that a remanence flux density of high-energy permanent magnet materials, such as neodymium-iron-boron, is more dependent on the operating temperature than conventional materials such as ferrites and alloys of iron, aluminum, nickel, copper and cobalt. Therefore, a significant degree of reversible demagnetization already occurs in the allowable temperature range. This has a corresponding effect on a concatenated permanent magnet flux from stator light and can be detected by means of a precise flux observer.

From the DE 10 2005 062 588 A1 a method for determining the magnet temperature of a permanently excited electric machine is known, wherein a phase voltage and a rotational speed of the electric machine is measured and used to determine the magnet temperature.

Furthermore, from the US 2014/0354204 A1 a method for determining a magnet temperature of a permanent-magnet electric motor, wherein a rotational movement of an engine and a status of a power supply of the motor are detected. A rotation speed of the motor is measured when it rotates with passive power supply. Depending on this rotation speed, the magnet temperature is estimated.

Further methods for determining a magnet temperature of a permanently excited electric motor are in A. Specht and J. Böcker: "Observer for the Rotor Temperature of IPMSM"; In: IEEE Power Electronics and Motion Control Conference, Ohrid, 2010 and A. Specht, O. Wallscheid and J. Böcker: "Determination of Rotor Temperature for an Interior Permanent Magnet Synchronous Machine Using a Precise Flux Observer"; In: IEEE International Power Electronics Conference, Hiroshima, 2014 described.

The invention is based on the object to provide a comparison with the prior art improved method for determining a magnet temperature of a permanently excited electrical machine.

The object is achieved by a method having the features specified in claim 1.

Advantageous embodiments of the invention are the subject of the dependent claims.

In the method according to the invention for determining a magnet temperature of a permanently excited electrical machine, at least one measured stator current of the machine, a stator voltage formed in a converter of the machine and a measured rotational speed of the machine are fed to a magnet temperature observing unit. A detection of the stator voltage at run time based on a model of the inverter, wherein by means of the Magnetettemperaturbeobachtereinheit from the supplied data and taking into account a flow temperature of a cooling medium of the machine, a permanent magnet flux is determined and by means of the magnet temperature observer unit from a flow difference between the permanent magnet flux and a reference magnet temperature associated and in the model stored in at least one map reference flux a temperature difference and from this the absolute magnet temperature is determined. To generate the at least one characteristic field, frequency-dependent influences on the machine are determined offline and imaged in the at least one characteristic diagram.

The inventive method allows due to the offline determination of the frequency-dependent influences on the machine, in particular in a systematic machine measurement, in a particularly advantageous manner that frequency-dependent influences of iron losses and frequency-dependent influences of measuring means, such as a rotary encoder or a current sensor, accurately determined and thus various error influences of the machine modeling and the measuring means can be taken into account in the model. Thus, a model accuracy is significantly increased. The term "offline" is understood to mean that the frequency-dependent influences on a test bench on the machine, ie. H. not be determined during normal operation of the machine, such as to drive a vehicle.

The model-based detection of the stator voltage at runtime on the basis of the model of an inverter allows in a simple manner without changing a construction of the machine, a cost-effective detection of the stator voltage. An actual measurement of the stator voltage is characterized by a significantly increased effort, required design changes and high costs.

The consideration of a flow temperature of a cooling medium of the machine makes it possible to compensate for fluctuations in an electrical-magnetic machine behavior and consequent influences on the magnetic temperature observation unit, particularly in the case of strongly fluctuating flow temperatures, for example in vehicle applications of the electric machine. Due to a thermal-mechanical expansion, different air gap widths occur at different temperatures. Due to a good thermal coupling of the cooling medium to the machine, the flow temperature of the cooling medium has a significant influence on the air gap width and thus on the electrical-magnetic machine behavior. Due to the consideration of the flow temperature of the cooling medium, the accuracy of the determination of the magnet temperature is further increased and estimation errors can be avoided.

Thus, it is possible by means of the method according to the invention to improve the model and to determine the magnet temperature of the electric machine very accurately. Due to the exact knowledge of the magnet temperature, improved and more accurate control and monitoring of the machine are possible. In particular, in this case an increase in thermal utilization of the electric machine can be realized, since due to the exact knowledge of the magnet temperature safety reserves can be reduced. Further, because a torque produced by the engine is very dependent on the magnet temperature, due to the accurate knowledge of the magnet temperature, it is possible to control the engine so that the torque generated by the engine can be set very accurately during its operation.

Also, errors can be detected earlier and avoided. In particular, due to the exact knowledge of the magnet temperature irreversible demagnetization of the permanent magnet of the machine can be avoided due to excessive operating temperatures.

Embodiments of the invention are explained in more detail below with reference to drawings.

Showing:

1 1 is a schematic view of a structure of a magnetic temperature observing unit for determining a magnet temperature of a permanent-magnet electric machine;

2 schematically a survey map of an electrical machine in a Anchorstellbereich and

3 schematically a survey map of an electrical machine in a Flußschwächbereich.

Corresponding parts are provided in all figures with the same reference numerals.

In 1 is a structure of a magnetic temperature observing unit 1 for determining a magnet temperature θ ^ _PM shown a permanently excited electric machine.

Precondition for a precise temperature determination is an exact observation of a chained permanent magnet _flux ψ ^ _dq [k] of the machine. For this purpose, the magnetic temperature observing unit 1 a measured stator current i _dq [k], a stator voltage u _dq [k] formed in a converter of the machine and, in a manner not shown, a measured speed of the machine.

The magnetic temperature observing unit 1 includes a model 2 and a feedback unit 3 wherein a basic structure of the magnetic temperature observing unit 1 one in A. Specht and J. Böcker: "Observer for the Rotor Temperature of IPMSM"; In: IEEE Power Electronics and Motion Control Conference, Ohrid, 2010 and A. Specht, O. Wallscheid and J. Böcker: "Determination of Rotor Temperature for an Interior Permanent Magnet Synchronous Machine Using a Precise Flux Observer"; In: IEEE International Power Electronics Conference, Hiroshima, 2014 corresponds structure described and based on an observation of the permanent _{magnetic flux} ψ ^ _dq [k] by means of a so-called reduced Luenberger structure. The magnet temperature θ ^ _PM is determined indirectly by means of a flow difference Δψ ^ _dq [k] between a permanent magnet _flux ψ ^ _dq [k] and a stored reference flux via a thermal material coefficient of a remanent flux density. This material coefficient is usually of the order of 0.1% flux density change per Kelvin. Even slight modeling and parameterization errors therefore lead to relatively large errors in the observed magnet temperature θ ^ _PM .

The model 2 a stator voltage u _dq [k] formed in a converter of the electrical machine is supplied as an input variable and in the manner shown this becomes based on a rotation matrix R (Δε [k]) taking into account an angle change Δε for several scanning indices [k] and a switching period T _{S determined} in an inverse Z-transformation z ^-1, a permanent magnet _flux ψ ^ _dq [k]. The permanent magnet _flux ψ ^ _dq [k] becomes one from the feedback unit 3 transmitted flow difference Δψ ^ _dq [k] between the permanent magnet _flux ψ ^ _dq [k] and one of a reference magnet _temperature θ _{PM, 0} assigned and in a map f -1 / dq of the model 2 stored reference _flow an estimate î _dq [k] of a stator current i _dq [k] generated as a controlled variable. This estimated value of the stator current î _dq [k] is multiplied within the model by an electrical stator resistance R _S and fed back via a rotation matrix R (Δε [k]) as a negative feedback.

(z = Operator) die Flussdifferenz Δψ ^_dq[k] erzeugt wird. Diese Flussdifferenz Δψ ^_dq[k] wird dem Modell 2 und einer zweiten Matrix

zugeführt, anhand welcher eine Temperaturdifferenz Δϑ ^_PM ermittelt wird. Aus dieser Temperaturdifferenz Δϑ ^_PM und der Referenzmagnettemperatur ϑ_PM,0 wird die Magnettemperatur ϑ ^_PM bestimmt.Furthermore, the estimated value of the stator current î _dq [k] of the feedback unit 3 as negative feedback to that of the feedback unit 3 supplied as an input variable stator current i _dq [k], wherein a difference of the stator current i _dq [k] and the estimate of the stator current î _dq [k] based on a first matrix

(z = operator) the flow difference Δψ ^ _dq [k] is generated. This flow difference Δψ ^ _dq [k] becomes the model 2 and a second matrix

supplied, based on which a temperature difference Δθ ^ _{PM is} determined. From this temperature difference Δθ ^ _PM and the reference magnet temperature θ _{PM, 0} , the magnet temperature θ ^ _{PM is} determined.

In particular, when using the permanently excited electric machine in partially electric or electrically driven vehicles, these are characterized by a very low power to weight ratio and power volume. This results in a high nonlinearity of electrical machine behavior, which z. B. caused by high iron saturation and cross-saturation. Furthermore, such machines are involved in gear integration, i. H. So-called gear-integrated machines, exposed to strong external temperature influences, which lead to a very wide operating temperature range. In this case, a stator and rotor temperature can differ greatly from each other.

• – Frequenzabhängige Einflüsse in der elektrischen Maschine, wie beispielsweise Wirbelstromverluste und Ummagnetisierungsverluste, und verwendeten Messmitteln, wie beispielsweise so genannten Winkelgebern und Strommesswandlern, die nicht in einem analytischen Maschinenmodell abgebildet werden können, verfälschen die Beobachtung des Permanentmagnetflusses ψ ^_dq[k] und führen zu signifikanten Abweichungen hinsichtlich der beobachteten Magnettemperatur ϑ ^_PM.
• – Eine wichtige Eingangsgröße der Magnettemperaturbeobachtereinheit 1 ist die Statorspannung u_dq[k]. Diese wird aus konstruktiven und kostentechnischen Gründen in Fahrzeugantrieben nicht gemessen und muss daher anderweitig indirekt ermittelt werden. Hierbei muss ein Einfluss des speisenden Umrichters, eine so genannte Umrichternichtlinearität, berücksichtigt werden, da regelungsinterne Sollspannungen nicht tatsächlichen Statorspannungen u_dq[k] entsprechen. Abweichungen in der Spannungsermittlung wirken sich direkt auf eine Genauigkeit der Magnettemperaturbeobachtung aus.
• – Aufgrund einer Wärmeausdehnung des Stators und Rotors können temperaturbedingte Änderungen einer Luftspaltbreite auftreten, welche wiederum zu einer Veränderung des beobachteten Permanentmagnetflusses ψ ^_dq[k] und damit zu einer Verfälschung des beobachteten Werts der Magnettemperatur ϑ ^_PM führen können.

Due to these design-specific features of permanently excited electrical machines in partially electric or electrically driven Vehicles encounter the following difficulties in modeling a electrical system with respect to permanent magnet observation:

• - Frequency-dependent influences in the electrical machine, such as eddy current losses and Ummagnetisierungsverluste, and used measuring means, such as so-called angle encoders and current transducers that can not be mapped in an analytical machine model, falsify the observation of the permanent magnet _flux ψ ^ _dq [k] and lead to significant deviations in the observed magnet temperature θ ^ _PM .
• An important input of the magnet temperature observing unit 1 is the stator voltage u _dq [k]. For reasons of design and cost, this is not measured in vehicle drives and therefore has to be determined indirectly indirectly. In this case, an influence of the supplying converter, a so-called inverter _nonlinearity , must be taken into account, since the _{controller's internal setpoint voltages} do not correspond to actual stator _voltages u _dq [k]. Deviations in the voltage determination have a direct effect on the accuracy of the magnetic temperature observation.
• Due to a thermal expansion of the stator and rotor temperature-induced changes of an air gap _width can occur, which in turn can lead to a change of the observed permanent _{magnetic flux} ψ ^ _dq [k] and thus to a falsification of the observed value of the magnet temperature θ ^ _PM .

Since, as described above in the first coat, an analytical description of the frequency-dependent influences of both iron losses and measuring means on the electrical-magnetic behavior of the machine under real-time conditions in vehicles is not possible or only with great effort, these influences are present by a systematic machine measurement, in particular at a test bench offline, determined and by means of at least one map f -1 / dq within the magnet temperature observing unit 1 displayed. The map f -1 / dq is formed in particular three-dimensional.

For this purpose, the machine is measured finely meshed in rotor-fixed coordinates, wherein for better interpolation additional operating points AP1 to APn outside a maximum current limit I _{max, motor of} a normal engine operation of the machine are approached to a maximum current limit I _{max, measuring} a motor measurement. The operating points AP1 to APn result from a transverse current i _q and a longitudinal current i _d . This survey map is recorded for different, constant speeds of the machine, with a speed resolution of about 1/10-steps relative to a maximum speed has proved advantageous.

2 shows a survey map of the electric machine thus created in a Anchorstellbereich, wherein a transverse current i _{q is shown} as a function of a longitudinal flow i _d .

In 3 a so created survey map of the electric machine is shown in a Flußschwächbereich, represent working points shown with a cross and a point AP1 to APn approached operating points and exclusively shown with a cross operating points AP1 to APn are target operating points of the machine.

For medium and high speeds, a voltage limit is set for the survey. For this purpose, a reduced operating point characteristic field in the stator current plane is specified, wherein a superimposed modulation controller displaces desired operating points, which lie outside the voltage limit, in the direction of a short-circuit point into a realizable working range during the measurement. As a result, the voltage limit for a given speed and a given DC link voltage is determined very accurately, which is also advantageous for an exact operating point control, d. H. Torque control of the machine, can be used.

When measuring the machine, each operating point AP1 to APn is approached in a stator current plane for a predetermined time, in particular a few seconds. The relevant current, voltage, torque, temperature and speed measurement variables as well as other measured variables are averaged over a period of time in a data preparation in order to achieve an ideal parameterization in the sense of a so-called fundamental wave model. In this case, all input variables of the magnetic temperature observer unit are also included 1 low pass filtered at runtime to realize fundamental wave modeling.

Because within the magnet temperature observing unit 1 the reference magnet temperature θ _{PM, 0 is} deposited, which correlates with a rotor temperature, this is already defined for the described measurement of the machine. In this case, the reference magnet temperature θ _{PM, 0 is selected} in particular from an area in the middle of a possible temperature range in order to achieve a rapid measurement at all rotational speeds. During the measurement, the rotor temperature is kept as constant as possible by using a superimposed temperature control during the measurement.

A flow temperature of a cooling medium of the machine has a significant influence on the measurement, since an air gap width can vary due to thermal expansion of the machine. The air gap range in turn has a significant influence on the magnetic-electrical machine behavior, which is identified during the measurement. Due to a good thermal coupling of the cooling medium to the machine, the flow temperature of the cooling medium is essentially responsible for the thermal expansion. Therefore, the flow temperature during the entire survey is kept constant. As a result, an at least substantially constant air gap width is ensured and influences of a varying flow temperature can be better compensated for at runtime. For the measurement, the flow temperature is selected, which corresponds to a time average of the flow temperature in the subsequent application of the machine, since thus the difference between the temperature during the measurement and in the application is minimized on average. Thus, the effects of a varying flow temperature are minimized.

Since the measurement of the machine takes place at a constant flow temperature, it is approximately assumed that the air gap width is constant during the measurement. Since the flow temperature during the application of the machine can vary significantly by a nominal value, the electrical-magnetic machine behavior also changes. For this reason, this influence is compensated in the magnetic temperature observing unit 1 ,

This compensation takes place in a possible embodiment by analytical adaptation. In this case, setting up a magnetic reluctance model in rotor-fixed coordinates allows an explicit consideration of the air gap width. This affects a magnetic coupling of rotor and stator with respect to the permanent magnet _flux ψ ^ _dq [k] as well as a stator self-induction. By further measuring the machine or by an analysis according to a finite element method with respect to a total possible flow temperature range, the influence on the measurement map and thus on the map f -1 / dq displayed. The survey map and thus the map f -1 / dq is then either a further dimension, namely the flow temperature, extended or manipulated by appropriate correction terms at the map output.

In another possible embodiment, the compensation is done by a so-called abstract fitting. Here, the fully parameterized magnet temperature observer unit becomes 1 operated during the measurement under varying flow temperatures. This results in a significant error in the observed magnet temperature θ ^ _PM , which is then used for the fitting. The input variable of the fitting is a temperature difference in the flow compared to the selected reference magnet temperature θ _{PM, 0} . The output represents a magnet temperature correction value corresponding to the output of the magnet temperature observing unit 1 , ie the magnet temperature θ ^ _PM , is added. The correction object used is a simple dynamic system model, for example a PT1 element, with additional polynomial correction. As a fitting approach, for example, the least squares method is used.

In addition, during the measurement of the machine, ie the determination of the frequency-dependent influences and the generation of the map f -1 / dq , a DC link voltage determined and taken into account. The DC link voltage determines the in 3 shown feasible operating points AP1 to APn of the machine in the Flußschwächbereich and is therefore kept constant during the survey. In traction applications of the machine, ie when used to drive a vehicle, a DC link is often provided by electrochemical energy stores or other sources of energy that provide a highly variable output voltage. This strongly variable output voltage results in accumulators, for example, from a Ladegradeinfluss. For the measurement of the machine, therefore, a maximum intermediate circuit voltage is selected based on a subsequent application of the machine plus a safety impact of, for example, 5%, so that all working points AP1 to APn in the flux weakening area are taken into account with a high level of safety.

Furthermore, a drive rate is determined and taken into account during the measurement of the machine. In applications of the machine for propulsion of a vehicle this is generally operated with a superimposed modulation controller and a fixed Aussteuergrad in Flußschwächbereich. Here, the Aussteuergrad is set constant during the survey and selected slightly higher than in the later real application.

In addition to the measurement of the machine, the stator voltage u _dq [k] is model-based at runtime, since an actual measurement leads to a very high cost and design _effort . A use of setpoint voltages from a machine control is not expedient in this case, since the influences of the supplying converter are not taken into account thereby. For realizing a real-time implementation, a mean value model of the converter is used, as it is for example in M. Seilmeier, Christoph Wolz and B. Piepenbreier: "Modeling and Model Based Compensation of Non-Ideal Characteristics of Two-Level Voltage Source Inverters for Drive Control Application", IEEE Electric Drives Production Conference, Nuremberg, 2011 is described. This modeling described there is expanded by determining and / or taking into account in the modeling an inverter temperature, line resistances and / or the intermediate circuit voltage.

Highly used drive inverters are subject to strong temperature fluctuations with regard to the power semiconductor modules. A semiconductor temperature has a strong influence on voltage drops in the forward direction of a transistor or a body diode. The semiconductor temperature also influences a transient switching and commutation behavior of the machine. For these reasons, the semiconductor temperature is taken into account in a converter measurement. If the inverter used does not have a temperature measuring device, a thermal model of the converter is used to determine the semiconductor temperature at runtime.

Since ohmic voltage drops occur in lines from the inverter to the machine, they are taken into account accordingly in converter modeling taking into account the line resistances.

Since, as already explained, the _{DC link} voltage can vary greatly, this is also used in the Umrichtermodellierung, so that in each control _cycle, the exact _{DC link} voltage for calculating the stator voltage u _dq [k] can be considered.

LIST OF REFERENCE NUMBERS

• 1 Magnet temperature observer unit
• 2 model
• 3 Feedback unit
• AP1 to APn operating point
• f -1 / dq map
• 
  first matrix
• 
  second matrix
• i _d longitudinal flow
• i _dq [k] stator current
• î _dq [k] _{Estimation of} the stator current
• I _{max, measuring} current limit of a motor _measurement
• I _{max, motor} current limit of a normal motor operation
• i _q cross flow
• [k] sample index
• T _S switching period
• u _dq [k] stator voltage
• R _S stator resistance
• R (Δε [k]) rotation matrix
• z variable of the z-transformation
• z ^-1 Inverse of the variable of the z-transformation
• Δε angle change
• Δθ ^ PM temperature difference
• Δψ ^ _dq [k] flow difference
• θ ^ _PM magnet temperature
• θ _{PM, 0} reference magnet _temperature
• ψ ^ _dq [k] permanent magnet flux

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 102005062588 A1 [0008]
• US 2014/0354204 A1 [0009]

Cited non-patent literature

• A. Specht and J. Böcker: "Observer for the Rotor Temperature of IPMSM"; In: IEEE Power Electronics and Motion Control Conference, Ohrid, 2010 [0010]
• A. Specht, O. Wallscheid and J. Böcker: "Determination of Rotor Temperature for an Interior Permanent Magnet Synchronous Machine Using a Precise Flux Observer"; In: IEEE International Power Electronics Conference, Hiroshima, 2014 [0010]
• A. Specht and J. Böcker: "Observer for the Rotor Temperature of IPMSM"; In: IEEE Power Electronics and Motion Control Conference, Ohrid, 2010 [0028]
• A. Specht, O. Wallscheid and J. Böcker: "Determination of Rotor Temperature for an Interior Permanent Magnet Synchronous Machine Using a Precise Flux Observer"; In: IEEE International Power Electronics Conference, Hiroshima, 2014 [0028]
• M. Seilmeier, Christoph Wolz and B. Piepenbreier: "Modeling and Model Based Compensation of Non-Ideal Characteristics of Two-Level Voltage Source Inverters for Drive Control Applications", IEEE Electric Drives Production Conference, Nuremberg, 2011 [0046]

Claims (5)

Method for determining a magnet temperature (θ ^ _PM ) of a permanently excited electric machine, - wherein a magnetic temperature observation unit ( 1 ) at least one measured stator current (i _dq [k]) of the machine, a stator voltage (u _dq [k]) formed in a converter of the machine and a measured speed of the machine are supplied, - wherein a detection of the stator voltage (u _dq [k ]) over a term based on a model ( 2 ) of the electric machine, - whereby by means of the magnetic temperature observation unit ( 1 ) is determined from the supplied data and taking into account a flow temperature of a cooling medium of the machine, a permanent magnet _flux (ψ ^ _dq [k]) and by means of the magnetic temperature _{observation unit} ( 1 ) from a flow difference (Δψ ^ _dq [k]) between the permanent magnet _flux (ψ ^ _dq [k]) and one associated with a reference magnet _temperature (θ _{PM, 0} ) and in the model ( 2 ) in at least one map (f -1 / dq) deposited reference flux a temperature difference (Δθ ^ _PM ) and from this the magnet temperature (θ ^ _PM ) is determined, wherein - to generate the at least one map (f -1 / dq) frequency-dependent influences on the machine are determined offline and in the at least one map (f -1 / dq) be imaged. A method according to claim 1, characterized in that in the generation of the at least one map (f -1 / dq) and in determining the magnet temperature (θ ^ PM) by means of the magnet temperature observing unit ( 1 ) A rotor temperature, the flow temperature of the cooling medium, a DC link voltage and a Ansteuergrad be determined and / or taken into account. A method according to claim 1 or 2, characterized in that when detecting the stator voltage (u _dq [k]) an inverter _temperature , line resistance and / or an intermediate circuit voltage is determined and / or taken into account. Method according to one of the preceding claims, characterized in that a change in the flow temperature of the cooling medium and a resulting change in an air gap width by means of a reluctance model in rotor-fixed coordinates, which is part of the model ( 2 ) is displayed. Method according to one of the preceding claims, characterized in that the at least one characteristic map (f -1 / dq) is determined at different speeds of the machine.

Images