DE102016220188A1 - Method for correcting measurement deviations of a sine-cosine rotation sensor - Google Patents

Abstract

Method (120) for correcting measurement deviations of a sine-cosine rotation sensor (100) for determining a rotational speed and / or an angular position of a rotating about a rotation axis member (106), wherein static errors of the sine-cosine rotation sensor (100) by a static correction (124) are compensated, characterized in that in addition a dynamic correction (122) takes place.

Description

• Aus der DE 10 2013 205 905 A1 ist ein Verfahren zur Bestimmung und/oder Ansteuerung einer Position eines Elektromotors, insbesondere in einem Kupplungsbetätigungssystem eines Kraftfahrzeuges, bekannt, bei welchem die Position eines Rotors des Elektromotors von einer, außerhalb einer Drehachse des Elektromotors an einem Stator des Elektromotors angeordneten Sensorik abgenommen wird, wobei das von der Sensorik abgenommene Positionssignal von einer Auswerteeinheit ausgewertet wird, wobei das während einer Sinusansteuerung des Elektromotors von der Sensorik abgegebene Positionssignal mittels mindestens eines, während einer Blockansteuerung des Elektromotors erfassten Positionssignals plausibilisiert wird.
• Aus der DE 10 2013 208 986 A1 ist ein Magnetgeberring einer Rotorlagesensorik eines elektrisch kommutierten Motors bekannt, welcher drehfest mit einem Rotor des elektrisch kommutierten Motors verbunden ist und welcher eine vorgegebene Anzahl von Magnetpolen mit einer alternierenden Magnetisierungsrichtung aufweist, wobei jedes Magnetpolpaar mindestens eine Einbuchtung aufweist.
• Aus der DE 10 2013 211 041 A1 ist ein Verfahren zur Bestimmung einer Position eines Elektromotors, insbesondere in einem Kupplungsbetätigungssystem eines Kraftfahrzeuges, bekannt, bei welchem ein Positionssignal eines Rotors des Elektromotors von einer, außerhalb einer Drehachse des Elektromotors an einem Stator des Elektromotors angeordneten Sensorik abgenommen wird und von einer Auswerteeinheit hinsichtlich der Position des Elektromotors ausgewertet wird, wobei nach einer Erkennung einer Änderung des Positionssignals eine Kommutierung einer Ansteuerung des Elektromotors ausgelöst wird, wobei nach der Erkennung der Änderung des Positionssignals eine Bestimmung der aktuellen Position des Rotors ausgeführt wird, wobei die Kommutierung des Elektromotors in Abhängigkeit von der detektierten aktuellen Position des Rotors ausgelöst wird.
• Aus der DE 10 2013 213 948 A1 ist ein Verfahren zur Bestimmung einer Position eines Elektromotors, insbesondere in einem Kupplungsbetätigungssystem eines Kraftfahrzeuges, bekannt, bei welchem ein Positionssignal eines Rotors des Elektromotors von einem, außerhalb einer Drehachse des Elektromotors an einem Stator des Elektromotors angeordneten Sensor abgenommen wird, welches von einer Auswerteeinheit hinsichtlich der Position des Elektromotors ausgewertet wird, wobei im Stillstand des Rotors dieser mit einer Spannung beaufschlagt wird und eine der Position des Rotors entsprechende Antwortreaktion einer Kommutierung des Elektromotors zugeordnet wird.
• Aus der DE 10 2013 222 366 A1 ist ein Verfahren zur Bestimmung und/oder Ansteuerung einer Position eines Elektromotors, insbesondere in einem Kupplungsbetätigungssystem eines Kraftfahrzeuges, bekannt, bei welchem die Position eines Rotors des Elektromotors von einer, außerhalb einer Drehachse des Elektromotors an einem Stator des Elektromotors angeordneten Sensorik abgenommen wird, wobei das von der Sensorik abgenommene Positionssignal von einer Auswerteeinheit ausgewertet wird, wobei das Positionssignal in Abhängigkeit von einer Übertragungsentfernung zwischen Sensorik und Auswerteeinheit bei kurzen Übertragungsentfernungen mittels eines SPI-Protokoll-Signals und/oder bei längeren Übertragungsentfernungen mittels eines PWM-Signals übertragen wird.
• Aus der DE 10 2012 220 629 A1 ist eine Rotorträgervorrichtung für eine einen Rotor und einen Stator aufweisende elektrische Maschine bekannt, mit einem auf einer Rotorwelle drehfest befestigbaren Rotorträgerelement, insbesondere drehfest befestigten Rotorträgerelement, an dessen Außenumfang der Rotor anordenbar ist, wobei ein axial am Rotorträgerelement angeordneter Blechring radial über den Außenumfang des Rotorträgerelements hinausragt und zu seiner dem Rotorträgerelement abgewandten Seite hin einen Kragen zur Aufnahme eines Resolvers der elektrischen Maschine ausbildet.
• Aus der DE 10 2012 205 024 A1 ist eine dynamoelektrische Maschine bekannt, aufweisend einen Stator, einem im Stator angeordneten und über einen Luftspalt von diesem beanstandeten Rotor, einem Resolver zur Bestimmung der Rotorlage und einem umformtechnisch hergestellten Resolverhalter zur Halterung eines Resolverstators des Resolvers, wobei der Resolver innerhalb des vom Rotor in axialer und radialer Richtung eingenommen Bauraums angeordnet ist.

The invention relates to a method for correcting measurement deviations of a sine-cosine rotation sensor for determining a rotational speed and / or an angular position of a rotating about an axis component, wherein static errors of the sine-cosine rotation sensor are compensated by a static correction.

• From the DE 10 2013 205 905 A1 is a method for determining and / or controlling a position of an electric motor, in particular in a clutch actuation system of a motor vehicle, known, in which the position of a rotor of the electric motor is removed from a, arranged outside a rotational axis of the electric motor to a stator of the electric motor sensor, wherein the position signal picked up by the sensor system is evaluated by an evaluation unit, wherein the position signal output by the sensor system during a sinusoidal control of the electric motor is made plausible by means of at least one position signal detected during a block control of the electric motor.
• From the DE 10 2013 208 986 A1 is a magnetic encoder ring rotor position sensor of an electrically commutated motor is known, which is rotatably connected to a rotor of the electrically commutated motor and which has a predetermined number of magnetic poles with an alternating direction of magnetization, each magnetic pole pair has at least one indentation.
• From the DE 10 2013 211 041 A1 is a method for determining a position of an electric motor, in particular in a clutch actuation system of a motor vehicle, known, in which a position signal of a rotor of the electric motor is removed from a, arranged outside a rotational axis of the electric motor to a stator of the electric motor sensor and by an evaluation unit with respect to Position of the electric motor is evaluated, wherein after detection of a change in the position signal, a commutation of a control of the electric motor is triggered, wherein after detecting the change of the position signal, a determination of the current position of the rotor is performed, wherein the commutation of the electric motor in dependence detected current position of the rotor is triggered.
• From the DE 10 2013 213 948 A1 is a method for determining a position of an electric motor, in particular in a clutch actuation system of a motor vehicle, known in which a position signal of a rotor of the electric motor is removed from a, arranged outside a rotational axis of the electric motor to a stator of the electric motor sensor, which from an evaluation unit in terms the position of the electric motor is evaluated, which is acted upon by a voltage at standstill of the rotor and a response of the position of the rotor corresponding response response of a commutation of the electric motor is assigned.
• From the DE 10 2013 222 366 A1 is a method for determining and / or controlling a position of an electric motor, in particular in a clutch actuation system of a motor vehicle, known, in which the position of a rotor of the electric motor is removed from a, arranged outside a rotational axis of the electric motor to a stator of the electric motor sensor, wherein the position signal picked up by the sensor system is evaluated by an evaluation unit, wherein the position signal is transmitted as a function of a transmission distance between sensor and evaluation unit at short transmission distances by means of an SPI protocol signal and / or at longer transmission distances by means of a PWM signal.
• From the DE 10 2012 220 629 A1 a rotor support device for a rotor and a stator having electrical machine is known, with a rotatably mounted on a rotor shaft rotor support member, in particular rotatably mounted rotor support member, on the outer circumference of the rotor can be arranged, wherein an axially on the rotor support member arranged sheet metal ring radially over the outer circumference of the rotor support member protrudes and forms a collar for receiving a resolver of the electric machine to its side facing away from the rotor support member side.
• From the DE 10 2012 205 024 A1 a dynamo-electric machine is known, comprising a stator, arranged in the stator and an air gap of the latter rotor, a resolver for determining the rotor position and a deformed Resolverhalter for holding a Resolverstators the resolver, wherein the resolver within the rotor in the axial and radial direction occupied space is arranged.

Due to the design and manufacturing variations of known from the prior art sensors as well as component scatters in the signal processing and analog-to-digital conversion arise angle-dependent and static deviations in the signals. With varying mechanical deformities of Sensor-stator to sensor-rotor as well as environmental and aging influences on the electronics also cause dynamic signal deviations during operation. This results in errors in the angular, speed and acceleration determination.

The invention has for its object to improve an aforementioned method for correcting errors in a sine-cosine rotation sensor.

The object is achieved with a method for the correction of measurement deviations of a sine-cosine rotation sensor having the features of claim 1. By virtue of the fact that a dynamic correction takes place in addition to the static correction, the correction of measurement deviations can be compared with those known from the prior art Continue to improve the process.

The dynamic correction can take place before the static correction. Preferably, the dynamic correction is taken into account in the subsequent static correction. The dynamic correction can be offset with the static correction. The dynamic correction can be subtracted from the static correction.

For static correction, a look-up table can be used. The lookup table can be created before starting the procedure. The look-up table can be created by means of a type and / or a piece test in comparison with a high-precision reference sensor.

During the procedure, exactly one value of the look-up table can be changed. During the procedure, several values of the look-up table can be changed. During the procedure all values of the look-up table can be changed. Measurement deviations that have already been compensated by the dynamic correction can be offset with the values of the look-up table. Measurement deviations that have already been compensated by the dynamic correction can be subtracted from the look-up table. In order to remove the measurement deviations from the look-up table, a least-squares estimation can be performed.

The rotating around the axis of rotation component whose speed and / or angular position is determined by the sine-cosine rotation sensor may be a rotor. The rotor may be a rotor of a dynamoelectric machine. The rotor may be a rotor of an electric motor. The rotor may be a rotor of a commutated electric motor. The rotor may be a rotor of an electric motor in a clutch actuation system of a motor vehicle.

The rotating about the axis of rotation component whose speed and / or angular position is determined by the sine-cosine rotation sensor may be part of an electrical machine. The electric machine may be a synchronous machine. The electric machine may be a brushless DC machine. The electric machine may have at least one phase. The electric machine may in particular have three phases. The electric machine can have at least two pole pairs. In particular, the electric machine can have four to fourteen pole pairs, in particular five pole pairs or eleven pole pairs. The three-phase windings can be controlled offset in time to form a moving drive field, which causes a drive torque on a permanent-magnet rotor. The rotor can move synchronously in operation to an AC voltage. The electric machine can be commutated depending on a movement position of the rotor, a movement speed of the rotor and / or a movement moment. A frequency and / or an amplitude can be variable depending on a movement position of the runner, a movement speed of the runner and / or a movement moment. The electric machine can be operated as a motor and / or as a generator. The electric machine may be for use in a motor vehicle. The electric machine can be used to drive a motor vehicle. The electric machine can be used to actuate a friction coupling device. The electric machine can be used to operate a hybrid disconnect clutch. The hybrid disconnect clutch may be for connecting an electric machine to or disconnecting an electric machine from a hybrid powertrain of a motor vehicle. The electric machine can be used to actuate a hydrostatic clutch actuator. The electric machine can be used to actuate a master cylinder of a hydrostatic clutch actuator. The electric machine can be used to actuate a transmission device. The electric machine may be a rotary machine in which the rotor is designed as a rotor and performs a rotary movement during operation. The rotor can be designed like a hollow shaft. The electric machine may be a linear machine in which the rotor performs a linear movement during operation. The electric machine can be block commutated. The electric machine may have a structurally integrated electrical control device. The Electric machine can be controlled by means of a structurally separate electrical control device. The electric machine may have at least one electrical power interface. The electrical machine may have at least one electrical signal interface. The electric machine can also be referred to as an actuator or actuator.

For parameter estimation, a high signal sampling rate can be used. For parameter estimation, a Newton method can be used. An estimate of the signal deviations can be made using Kalman filters. An estimate of the signal deviations can be made by means of neural networks.

The sine-cosine rotation sensor may be formed as a resolver. A resolver is an electromagnetic transducer for converting the angular position of a rotor into an electrical variable. The sine-cosine rotation sensor can operate on the basis of the AMR effect (anisotropic magnetoresistive effect). The sine-cosine rotation sensor can operate on the basis of the GMR (giant magnetoresistor) effect. The sine-cosine rotation sensor can operate based on the TMR (magnetic tunnel resistance) effect. In summary and in other words, the invention thus provides, inter alia, a method for increasing the measuring accuracy of sine-cosine rotation sensors, in particular resolvers. As a rule, static and dynamic signal deviations occur together in sine-cosine rotation sensors. If only one or the other is compensated, an uncompensated error always remains. This is, for example, one from the publication "Q. Lin, T. Li, Z. Zhou: Error Analysis and Compensation of the Orthogonal Magnetic Encoder, 2011 International Conference on Instrumentation, Measurement, Computer, Communication and Control " known method, since the signal deviations are determined only once, and are statically corrected in operation, with dynamic signal deviations in this solution continue to lead to errors.

According to the invention, a combination of off-line and online-based correction is proposed to increase the measurement accuracy of sine-cosine rotation sensors. This can be used to compensate both static and dynamic deviations. For offline-based compensation, a simple look-up table is proposed, which is described by means of a type or a piece test in comparison with a high-precision reference sensor. In this off-line compensation, it is important that the portions of the dynamic compensation be removed from the look-up table. This is based on the idea that these dynamic components are compensated online. In the software, an estimate of the signal deviations and their compensation are switched before this offline look-up table. This compensates for dynamic changes in operation. In order to be able to follow fast temporal variations, a high signal sampling rate and the Newton method for parameter estimation are proposed. By alternative methods of estimating the signal deviations, comparatively fast parameter convergences could be achieved, for example by means of Kalman filters or neural networks.

The invention can be used in all angular speed and acceleration measurements with high accuracy requirements. An application of the described adjustment method is conceivable in the field of electric drives. This results in advantages in the sensor costs as well as in the efficiency and the control quality.

The invention provides a method for correcting measurement deviations of sine-cosine rotation sensors, which enables a compensation of static and dynamic signal deviations to increase the measurement accuracy. In particular, a combined adjustment method of sine-cosine rotation sensors is provided.

Hereinafter, an embodiment of the invention will be described with reference to figures. From this description, further features and advantages. Concrete features of this embodiment may represent general features of the invention. Features associated with other features of this embodiment may also represent individual features of the invention.

• 1 eine schematische Darstellung eines aus dem Stand der Technik bekannten Resolvers,
• 2 eine schematische Darstellung eines erfindungsgemäßen Verfahrens zur Korrektur von Messabweichungen eines Resolvers, und
• 3 Winkelabweichungen unterschiedlich korrigierter Messabweichungen des Resolvers.

They show schematically and by way of example:

• 1 a schematic representation of a known from the prior art resolver,
• 2 a schematic representation of a method according to the invention for the correction of errors of a resolver, and
• 3 Angular deviations of differently corrected measuring deviations of the resolver.

1 shows a known from the prior art sine-cosine rotation sensor, also called a quadrature sensor, which in the present case as a resolver 100 is executed. Such a resolver 100 is known, for example, from the data sheet "AD2S1210, Variable Resolution, 10-Bit to 16-Bit R / D Converter with Reference Oscillator" from the year 2010.

mit E₀: Anregungsamplitude, ω: Anregungsfrequenz.A first stator winding 102 comes with a sinusoidal first AC voltage 104 applied. The first alternating voltage 104 is calculated as a function of time t $$e_0\cdot sin\left(\omega t\right)$$

with E ₀ : excitation amplitude, ω: excitation frequency.

A rotor 106 points opposite to a reference angle position 108 a rotation angle 110 (Angle of rotation α).

A second stator winding 112 lies on one of the first stator windings 102 opposite side of the rotor 106 , A third stator winding 114 is opposite to the second stator winding 112 arranged rotated by 90 degrees.

mit E₀: Anregungsamplitude, ω: Anregungsfrequenz, T: Transformationsverhältnis.The first alternating voltage 104 energized in the second stator winding 112 a second AC voltage 116 , The second AC voltage 116 is calculated as a function of the time t and the angle of rotation 110 (Angle of rotation α) to: $$e_0\cdot T\cdot sin\left(\omega t\right)\cdot cOs\left(\alpha \right)$$

with E ₀ : excitation amplitude, ω: excitation frequency, T: transformation ratio.

The second AC voltage 116 is the cosine of the rotation angle 110 (Angle of rotation α) modulated.

mit E₀: Anregungsamplitude, ω: Anregungsfrequenz, T: Transformationsverhältnis.The first alternating voltage 104 energized in the third stator winding 114 a third alternating voltage 118 , The third alternating voltage 118 is calculated as a function of the time t and the angle of rotation 110 (Angle of rotation α) to: $$e_0\cdot T\cdot sin\left(\omega t\right)\cdot sin\left(\alpha \right)$$

with E ₀ : excitation amplitude, ω: excitation frequency, T: transformation ratio.

The third alternating voltage 118 is the sine of the rotation angle 110 (Angle of rotation α) modulated.

The amplitudes of the second alternating voltage 116 and the third AC voltage 118 are thus dependent on the angle of rotation 110 (Angle of rotation α) of the rotor 106 , One between the resolver 100 and a system microprocessor arranged resolver-to-digital converter (analog-to-digital converter) uses the sine and cosine signals of the second and third AC voltages 116 . 118 for decoding the angular position and speed of the rotor 106 ,

Due to the design and manufacturing variations of sine-cosine rotation sensors, in this case the resolver 100 , as well as component scatters in the signal conditioning and the analog-to-digital conversion arise angle-dependent and static deviations in the signals. With varying mechanical misalignments from stator to rotor 106 as well as environmental and aging influences on the electronics also cause dynamic signal deviations during operation.

2 schematically shows a method according to the invention 120 for correction of measuring deviations of a sine-cosine rotation sensor using the example of the resolver 100 , The resolver 100 delivers during the procedure 120 as analog output signals, the second AC voltage 116 and the third AC voltage 118 ,

In one process step 122 finds a dynamic correction of the two output signals second AC voltage 116 and third AC voltage 118 instead of. The dynamic correction 122 causes a Compensation during the procedure 120 determined online or online estimated amplitude deviations and / or offset deviations of the second AC voltage 116 and the third AC voltage 118 ,

In one of the dynamic correction 122 downstream process step 124 finds a static correction of the two output signals second AC voltage 116 and third AC voltage 118 instead of. Thereby a compensation of already before the beginning of the procedure takes place 120 determined static deviations to a high-precision reference sensor, not shown in the figures instead. The previously made dynamic correction 122 the output signals will be in static correction 124 taken into account, in particular deducted.

To compensate for the static measurement deviations, a look-up table is stored in the software. The look-up table contains information during the procedure 120 be used to avoid complicated calculations. The look-up table will appear before the start of the procedure 120 based on individual end-of-line measurements, ie measurements in the final assembly area of the resolver 100 , or based on typical values for the sensor series of the resolver 100 certainly. Using the look-up table, you can use interpolation for each angle of rotation 110 (Rotation angle α) the necessary static correction can be calculated.

During the static correction 124 In the look-up table, the measurement deviations are deducted, which were previously due to the dynamic correction 122 were compensated. These include in particular amplitude and offset deviations in the sine and cosine signals of the two output signals second AC voltage 116 and third AC voltage 118 , Through the dynamic correction 122 and the subsequent static correction 124 results in a corrected measurement result 126 ,

During the procedure 120 the described dynamic compensation is performed by the Newton method and then the static compensation.

$$y_{s,k}=a_s\cdot \text{sin}\left({\alpha }_k\right)+b_s$$

To compensate for the parameter fluctuations, the following model of the optionally demodulated output signals is assumed. The individual signal amplitudes of cosine and sine (second AC voltage 116 , third alternating voltage 118 ) are described by the coefficients a, the offsets by b. A time index is marked k. $$y_{c.k}=a_c\cdot \text{cos}\left({\alpha }_k\right)+b_c$$

$$y_{s.k}=a_s\cdot \text{sin}\left({\alpha }_k\right)+b_s$$

In order to remove the measurement deviations resulting from amplitude and offset deviations from the static look-up table, it is possible to carry out a least-squares estimation of these parameters and thus of the measurement deviations caused thereby. For this, the two equations above are squared and added: $$1=\frac{1}{a_c^2}\cdot y_{c.k}^2-\frac{b_c}{a_c^2}\cdot y_{c.k}+\frac{1}{a_s^2}\cdot y_{s.k}^2-\frac{b_s}{a_s^2}\cdot y_{s.k}+\frac{b_s^2}{a_s^2}+\frac{b_c^2}{a_c^2}$$

$${\begin{array}{cccc}\underset{\_}{\theta }=[\frac{1}{a_c^2}& \frac{1}{a_s^2}& -\frac{b_c}{a_c^2}& -\frac{b_s}{a_s^2}]\end{array}}^T$$

$$\begin{array}{cc}\underset{\_}{1}=\underset{\_}{Y}\cdot \underset{\_}{\theta }& |\frac{b_s^2}{a_s^2}+\end{array}\frac{b_c^2}{a_c^2}\ll 1$$

The associated measurement and parameter vector and the simplified vector equation are: $$\begin{array}{cccc}\underset{\_}{Y}=[y_{c.k}^2& y_{s.k}^2& y_{c.k}& y_{s.k}]\end{array}$$

$${\begin{array}{cccc}\underset{\_}{\theta }=[\frac{1}{a_c^2}& \frac{1}{a_s^2}& -\frac{b_c}{a_c^2}& -\frac{b_s}{a_s^2}]\end{array}}^T$$

$$\begin{array}{cc}\underset{\_}{1}=\underset{\_}{Y}\cdot \underset{\_}{\theta }& |\frac{b_s^2}{a_s^2}+\end{array}\frac{b_c^2}{a_c^2}"1$$

The signal deviation parameters can thus be determined on the basis of a measurement series as follows. $$\widehat{\theta }={\left(Y\text{'}\cdot Y\right)}^{-1}\cdot Y\text{'}\cdot \underset{\_}{1}$$

These parameters allow the determination of the associated dynamic angle-dependent measurement deviations. These are deducted from the measurement deviations to a high-precision reference sensor and the result is stored in the static look-up table.

The system of equations in matrix representation for dynamic online compensation is: $$\left[\begin{array}{c}\text{cos}{\alpha }_k\\ \text{sin}{\alpha }_k\end{array}\right]=\underset{M_k}{\underset{\}}{\left[\begin{array}{r}\hfill \begin{array}{cccc}{y}_{c.\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE102016220188A1\_0018.png"\; state="\{\{state\}\}">k</figure-callout>}}& 1& 0& 0\end{array}\\ \hfill \begin{array}{cccc}0& 0& y_{s.k}& 1\end{array}\end{array}\right]}}\cdot \underset{\underset{\_}{\theta }}{\underset{\}}{{\left[\begin{array}{cccc}\frac{1}{a_c}& -\frac{b_c}{a_c}& \frac{1}{a_s}& -\frac{b_s}{a_s}\end{array}\right]}^T}}$$

$$J={\displaystyle \sum _{k=0}^{N-1}{\left({\Vert M_k\cdot \underset{\_}{\theta }\Vert }_2^2-1\right)}^2}$$

Based on the square of the L2 standard, the merit function J is defined: $$\underset{\_}{1}={\Vert M\cdot \underset{\_}{\theta }\Vert }_2^2$$

$$J={\displaystyle \underset{k=0}{\overset{N-1}{\Sigma }}{\left({\Vert M_k\cdot \underset{\_}{\theta }\Vert }_2^2-1\right)}^2}$$

The derivation of the quality function J according to the parameter vector theta results in: $$\nabla J={\displaystyle \underset{k=0}{\overset{N-1}{\Sigma }}4\cdot M_k^T\cdot M_k\cdot \underset{\_}{\theta }\cdot }\left({\underset{\_}{\theta }}^T\cdot M_k^T\cdot M_k\cdot \underset{\_}{\theta }-1\right)$$

For the Newton method, moreover, the Hesse matrix H is needed, ie the second derivative of the quality function J according to the parameter vector: $$H={\displaystyle \underset{k=0}{\overset{N-1}{\Sigma }}4\cdot M_k^T\cdot M_k\cdot (I\cdot }\left({\underset{\_}{\theta }}^T\cdot M_k^T\cdot M_k\cdot \underset{\_}{\theta }\right)+2\cdot \underset{\_}{\theta }\cdot {\underset{\_}{\theta }}^T\cdot M_k^T\cdot M_k)$$

The gradient and the Hesse matrix are calculated iteratively. To do this, the argument of the sum function is determined in each time step and added to the scaled previous value. The scaling factor is referred to as the forgetting factor.

With a step size gamma, the signal parameters can be estimated iteratively. The step size can be adapted, for example by means of so-called backtracking. Due to a high sampling rate of the sensor signals, a fast convergence of this parameter estimation can be ensured. $$\underset{\_}{\theta }\to \underset{\_}{\theta }-\gamma \cdot H^{-1}\cdot \nabla J$$

In 3 the results of the correction in a diagram are exemplary for the resolver 100 shown. The angle deviations are over an electrical angle 126 represented, the present the angle of rotation 110 of the rotor 106 equivalent. The value of an uncorrected angular deviation 128 varies over an electrical angle 126 of 360 °, in particular one revolution of the rotor 106 corresponds to a minimum uncorrected angular deviation 130 and a maximum uncorrected angular deviation 132 , The value of a statically corrected angular deviation 134 varies over the electrical angle 126 of 360 ° between a minimum statically corrected angular deviation 136 and a maximum statically corrected angular deviation 138 , The value of a dynamically corrected angular deviation 140 varies over the electrical angle 126 of 360 ° between a minimum dynamically corrected angular deviation 142 and a maximum dynamically corrected angular deviation 144 , The value of a statically and dynamically corrected angle deviation 146 varies over the electrical angle of 360 ° between a minimum statically and dynamically corrected angular deviation 148 and a maximum statically and dynamically corrected angular deviation 150 ,

In the in 3 illustrated example, the maximum deviations of the electrical angle 126 without correction -1.29 ° to 1.32 °, with static correction -1.04 ° to 1.09 °, with dynamic correction -0.608 to 0.651 ° and with static and dynamic correction -0.488 ° to 0.447 °. The measuring accuracy can be determined with the aid of the method according to the invention 120 thus improve by almost a factor of three.

The described invention is equally applicable to other sine-cosine rotation sensors operating on the basis of, for example, the AMR effect (anisotropic magnetoresistance effect) and / or GMR effect (giant magnetoresistance) and / or TMR effect (magnetic tunnel resistance).

LIST OF REFERENCE NUMBERS

100

resolver

102

first stator winding

104

first alternating voltage

106

rotor

108

Reference angular position

110

angle of rotation

112

second stator winding

114

third stator winding

116

second AC voltage

118

third alternating voltage

120

method

122

Procedural step, dynamic correction

124

Procedural step, static correction

126

electrical angle

128

uncorrected angular deviation

130

minimal uncorrected angular deviation

132

maximum uncorrected angular deviation

134

statically corrected angular deviation

136

minimal statically corrected angular deviation

138

maximum statically corrected angular deviation

140

dynamically corrected angular deviation

142

minimum dynamically corrected angle deviation

144

maximum dynamically corrected angle deviation

146

statically and dynamically corrected angle deviation

148

minimal statically and dynamically corrected angular deviation

150

maximum statically and dynamically corrected angular deviation

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 102013205905 A1 [0001]
• DE 102013208986 A1 [0001]
• DE 102013211041 A1 [0001]
• DE 102013213948 A1 [0001]
• DE 102013222366 A1 [0001]
• DE 102012220629 A1 [0001]
• DE 102012205024 A1 [0001]

Cited non-patent literature

• "Q. Lin, T. Li, Z. Zhou: Error Analysis and Compensation of the Orthogonal Magnetic Encoder, 2011 International Conference on Instrumentation, Measurement, Computer, Communication and Control "[0011]

Claims (10)

Method (120) for correcting measurement deviations of a sine-cosine rotation sensor (100) for determining a rotational speed and / or an angular position of a rotating about a rotation axis member (106), wherein static errors of the sine-cosine rotation sensor (100) by a static correction (124) are compensated, characterized in that in addition a dynamic correction (122) takes place. Method (120) according to Claim 1 , characterized in that the dynamic correction (122) takes place before the static correction (124). Method (120) according to at least one of the preceding claims, characterized in that the dynamic correction (122) is taken into account in the subsequent static correction (124), in particular subtracted from the static correction (124). Method (120) according to at least one of the preceding claims, characterized in that a look-up table is used for the static correction (124), which was created in particular before the beginning of the method (120). Method (120) according to Claim 4 , characterized in that during the method (120) at least one value of the look-up table is changed, in particular measurement deviations that have already been compensated by the dynamic correction (122) are subtracted from the look-up table. Method (120) according to at least one of the preceding claims, characterized in that a least-squares estimation is performed in order to remove measurement deviations from the look-up table. Method (120) according to at least one of the preceding claims, characterized in that the component rotating about the axis of rotation is a rotor (106) of an electrical machine. Method (120) according to at least one of the preceding claims, characterized in that a high signal sampling rate and / or a Newton method for parameter estimation is used. Method (120) according to at least one of the preceding claims, characterized in that an estimation of the signal deviations takes place by means of Kalman filters and / or neural networks. Method (120) according to at least one of the preceding claims, characterized in that the sine-cosine rotation sensor is designed as a resolver (100).

Images