DE10226974A1 - Method for determining rotor position of field-orientated driven polyphase machine, involves determining test signal conditioned current AC component each switching period depending on ascertained current amplitudes - Google Patents

Abstract

A method for determining rotor position (16) involves providing a test space vector which is synchronous with the pulse-width modulation of a field-orientated control of the polyphase machine then component-wise superposing of this test-space vector to the field orientated components of a desired voltage vector synchronous to the pulse-width modulation. At least two current amplitudes are determined per switching period (Ts) of a current space vector of the machine. A test signal conditioned current AC component (Aw) is converted into an error angle (e) and estimated rotor position is corrected. An independent claim is included for a device for the rotor position determination method.

Description

The invention relates to a Method for determining a rotor position of a field-oriented one Three-phase machine, which has an effective inductance dependent on the rotor position, without a mechanical sensor at standstill and at low speed and an apparatus for performing this method.

• 1. Bestimmung der Rotorlage für eine feldorientierte Regelung. Das bedeutet, dass der Strom im Motor immer entsprechend dem Rotorwinkel eingestellt wird und so exakt das geforderte Moment eingestellt werden kann.
• 2. Bestimmung der Lastposition aus der Rotorlage. Die Lage der über eine feste Übersetzung an dem Motor angekoppelten Last wird aus der Rotorlage bestimmt. Dazu wird beim Start des Systems ein Referenzpunkt angefahren und dann im Betrieb die Lage mitgeführt.

A permanent magnet synchronous machine is mostly used for servo drives. Thanks to their commutatorless design, the machines are low-maintenance and their high power-to-weight ratio leads to a small construction volume. The field-oriented control of the permanent-magnet synchronous machine decouples the control of the flux and torque-forming current components. For this purpose, the currents are transformed to a rotor flux-oriented coordinate system. Since the rotor flux of the synchronous machine is constant in terms of amount and stationary with respect to the rotor, the rotor flux angle required for the transformation can easily be determined using a mechanical sensor. An optical encoder or a magnetic resolver is used as the mechanical sensor. The mechanical sensor usually performs two functions:

• 1. Determination of the rotor position for a field-oriented control. This means that the current in the motor is always set according to the rotor angle and the required torque can be set exactly.
• 2. Determination of the load position from the rotor position. The position of the load coupled to the motor via a fixed ratio is determined from the rotor position. This is done at the start approached a reference point of the system and then carried the situation during operation.

The rotor position measurement, however, causes noteworthy Costs because one encoder, one slot, one connector, one Encoder line and an encoder evaluation must be provided.

An encoderless, field-oriented regulation will previously only used with asynchronous motors. However, this works only from a minimum speed, for example a few percent of Nominal speed, so it is not suitable for positioning. It also says no rotor position information needed to determine the position of the load could be used for Available.

A similar donorless, field-oriented operation is also available with the permanently excited synchronous machine. It is based on the voltage induced during rotation and is therefore also the first possible from a minimum speed and therefore not suitable for positioning.

WO 92/01331 A1 describes a method and a circuit arrangement for sensorless rotation angle detection of a damper-less, preferably permanent magnet excited, powered by a converter Synchronous machine known. This known method uses block-shaped, stator-oriented Test vectors. It breaks the current control loop by taking place of the modulator output applies a sequence of short test pulses. there the current is always at the full DC link voltage and dismantled again. This process is carried out in succession for all three phases. The direct transition from an active pointer to the opposite active pointer causes an increased Voltage jump (double DC link voltage), which may be unfavorable for the motor insulation is. The current change caused by the voltage pulses is done by quickly sensing the current during the current rise and while of the current drop measured. So it will be a fast one and synchronized current measurement is required. Because the location of the current pulses is determined by the stator windings and accordingly depending on the rotor position high momentum pulses arise which leads Test pulses for high noise. This procedure requires interventions in the pulse generation and in the current regulation. Moreover is location information only available after the test blocks have been introduced, what corresponds to a limitation of the bandwidth of the position detection.

From the EP 0 228 535 A1 A method and a device for determining the flow angle of a induction machine or for position-oriented operation of the machine are known. In this method, a high-frequency component is impressed on an electrical state variable of the stator winding system. The electrical state variables of the stator winding system are understood to mean the currents and voltages in the individual stator coils and the axes of those coils to which the high-frequency component is impressed determine the direction of the impressed high-frequency component. It turns out that, for example, in the case of an impressed high-frequency current in one or more stator coils, high-frequency components also occur in the voltage of these coils and in the currents and voltages of the other coils, the amplitude of which depends on the difference angle between the flow axis and the direction of the impressed component. Therefore, the amplitude of the high-frequency component is detected from a state signal that represents another state variable of the stator winding system; the directional angle sought is determined from the dependence of the detected amplitude on the predetermined direction of the impressed high-frequency part.

According to this method, it is now provided to add a high-frequency component to the stator current by means of an additional target vector that is high-frequency compared to the frequency of the reference vector overlap. The superimposition can be carried out by vector addition to the field-oriented target vector or to the corresponding vector transferred into the stand-oriented coordinate system or in another, mathematically equivalent manner. Consequently, the control vector also receives a correspondingly high-frequency component, which is impressed on the stator current via the converter. If the additional vector is a vector with a rapidly changing direction, the control can also be changed correspondingly quickly. By embossing a high-frequency component in the stator current or in the stator voltage, high-frequency voltage components (or current components) are coupled into the stator winding, the envelopes of which are assigned to the position of the field or rotor axis. These envelopes are assigned to the components of the direction vector by suitable means. A flow computer is provided to form the direction vector, which calculates the stand-related components of a model vector describing the flow from electrical quantities of the machine.

This method is based on the observation that one into one of the stator coils embossed high frequency Proportion of one state variable high-frequency component of the other state variable of the same coil and high-frequency components of the state variables of other coils induced. The high frequency components depend on the position of the runners or field axis. A sinusoidal oscillation is used as an additional target vector of about 250 Hz. Using this method is a field-oriented Operation of a synchronous machine possible.

From US-PS 6 005 365 is one Engine control device for a synchronous machine without a mechanical sensor is known. At this Motor control device also uses a sinusoidal test signal that overlaps in the d-axis and is specified as the current setpoint. The is evaluated Cross coupling.

Similar also works the process that is described in the article entitled "Rotor Position and Velocity Estimation for a Salient-Pole Permanent Magnet Synchronous Machine at Standstill and High Speeds ", released in the magazine “IEEE Transactions on Industry Applications ", Vol. 34, No. 4, July / Aug. 1998, pages 784 to 789. However, this procedure is included consideration the speed-related cross coupling and with an observer structure for the Frequency control equipped. In addition, a synchronous Rectification used to determine the cross component of the current. In this method, however, the test signal is in the torque-forming Axis laid. This causes a torque pulsating with the test frequency, which leads to whistling and mechanical vibrations. For that results there is a better decoupling of the position detection from the current control loop.

A disadvantage of all of these methods the small usable width, since the test signal is significantly lower in frequency than the switching frequency must be and the usable bandwidth again is significantly below the test signal frequency.

The object of the invention is now is based on a method for determining a rotor position of a field-oriented operated induction machine, which is an effective depending on the rotor position inductance has to indicate at which the bandwidth is significantly increased.

This object is achieved with the Process steps of claim 1 solved.

By synchronizing the test space pointer to the pulse width modulation of a field-oriented control of the Rotary field machine with an effective inductance dependent on the rotor position achieved a high usable bandwidth. Furthermore arise at Determination of the rotor position less noise. In addition, this synchronicity of the test space pointer with pulse width modulation a clean decoupling of the field-oriented control required voltage space vector and of the test space pointer.

This is a synchronous test space pointer for pulse width modulation generates a test signal-related in the current space pointer of the induction machine Current change component, which is only from the test signal and the rotor position of the field-oriented operated induction machine. Besides, is this alternating current share independently from the requested voltage space vector. This is the share of the current change caused by the test signal proportional to an error angle between an estimated and the actual Rotor position occurs. As soon as this error angle is regulated to zero the estimated rotor position corresponds to the actual one Rotor position of the field-oriented induction machine. For investigation of this test-related current change component will be at least two Current amplitudes used per switching period of a current space vector. Since the test space pointer is synchronous with the pulse width modulation, can with simple means the test-related share of electricity changes again from the motor current space vector of the field-oriented induction machine be separated. This is a complete decoupling of the test space pointer from the current control of the field-oriented control of the induction machine possible.

In an advantageous method, the direction of the test space pointer is adjusted to the determined rotor position in such a way that it lies in the d-axis of the field-oriented induction machine. As a result, no torque is generated by a test current space pointer caused by the test space pointer. This will reduce the noise level Rotary field machine in sensorless rotor position determination significantly reduced.

Further advantageous configurations of the method according to the invention are the dependent claims 3 to 6.

To further explain the invention, reference is made to the drawing referenced in which a device for performing the inventive method is illustrated schematically.

1 shows the structure of a known servo control in which

2 is shown a servo control with a device for performing the method according to the invention

3 shows a cross section through a servo motor, whereas in the

4 the coordinate systems are shown in a sensorless rotor position determination in which

5 several signal curves are each shown in a diagram over time t, which reflects the method according to the invention, and in

6 are several voltage profiles each with an associated current profile each illustrated in a diagram over the time t, the

7 based on the evaluation of several current measured values per switching period, the complete decoupling of the test space pointer from the current control is shown.

According to the structure of the known servo control, it consists of a position control loop 2 , a speed control loop 4 , a current control loop 6 and a transformation device 8th , This control generates a room pointer as a manipulated variable, which is generated using a modulator 10 in control signals for a power unit 12 , for example a self-commutated pulse converter, is converted. These control signals are by means of a locking device 14 locked together. The power section 12 controls a three-phase machine 16 with an effective inductance dependent on the rotor position Θ. Such an induction machine 16 is a permanently excited synchronous machine. The current control loop 6 is the speed control loop 4 subordinate to the position control loop 2 is subordinate.

The position control loop 2 consists of a controller 20 and a comparator 22 , the output side with the input of the controller 20 is linked. At the inverting input of the comparator 22 there is a determined rotor position value T, whereas an associated setpoint Θ * at the non-inverting input of this comparator 22 pending. At the output of the controller 20 of the position control loop 2 there is a speed control variable Θ , _y to a non-inverting input of a comparator 24 of the subordinate speed control loop 4 is fed. This speed control loop too 4 has a regulator 26 on that of the comparator 24 is connected downstream. At the inverting input of the comparator 24 there is a determined actual speed value Θ , which is derived, for example, from the rotor position value T by differentiation. At a second non-inverting input there is a speed setpoint #. .* on. Using this speed control loop 4 a setpoint of a torque-forming current component i * / q is generated. Because the induction machine 16 is a permanently excited synchronous machine, no flux control loop is required that generates a flux-forming current setpoint component i * / d. For this reason, the setpoint of the flux-forming current component i * / d is zero.

. Diese feldorientierten Spannungskomponenten u * / q und u * / d werden mittels eines Vektordrehers 36 der Transformationseinrichtung 8 in ständerorientierte Spannungskomponente u * / und u * / α gedreht. Für diese Vektordrehung wird der ermittelte Rotorlagewert Θ benötigt. Mittels eines weiteren Vektordrehers 38 dieser Transformationseinrichtung 8 werden auch die ermittelten ständerorientierten Stromkomponenten i_α und i_β des Motorstroms-Raumzeigers

in feldorien tierte Stromkomponenten i_d und i_q unter Zuhilfenahme des ermittelten Rotorlagewertes Θ gedreht. Diese Transformationseinrichtung 8 ist somit eine Schnittstelle zwischen der feldorientierten Regelung und der stromrichtergespeisten Drehfeldmaschine 16.The current control loop 6 This field-oriented control has two control channels for the two field-oriented current components i * / d and i * / q. Each control channel has a comparator 28 respectively. 30 and a regulator 32 respectively. 34 on. At the non-inverting inputs of these comparators 28 and 30 there is a setpoint, the torque and flux-forming current components i * / q and i * / d. The corresponding actual values of the torque and flux-forming current components i _q and i _{d are present} at their inverting inputs. Any regulator 32 and 34 generates a voltage value of a voltage component u * / q and u * / d of the current control loop 6 required voltage space vector

, These field-oriented voltage components u * / q and u * / d are generated using a vector rotator 36 the transformation device 8th into stand-oriented voltage component u * / and u * / α turned. The determined rotor position value Θ is required for this vector rotation. Using another vector rotator 38 this transformation device 8th also the determined stator-oriented current components i _α and i _{β of} the motor current space vector

rotated into field-oriented current components i _d and i _q with the help of the determined rotor position value Θ. This transformation device 8th is thus an interface between the field-oriented control and the converter-fed induction machine 16 ,

So that the generated voltage setpoint pointer u * / d and u * / q by means of the power section 12 to the induction machine 16 can be applied, the two stand-oriented voltage components u * / α and u * / β by means of the modulator 10 , for example a pulse width modulator, are converted into phase-related control signals. These phase-related control signals are generated by means of the locking device 14 locked against each other. This locking device 14 ensures that there is no short circuit in the power section 12 can occur with functional semiconductor switches. The one in the induction machine 16 flowing current is recorded in phases and by means of a coordinate converter 40 transformed into orthogonal current components i _α and i _{β of} a column-oriented Cartesian coordinate system.

anstehen. Außerdem weist diese Struktur der bekannten Servoregelung eine Vorrichtung 44 auf, mit der in bekannter Weise die Rotorlage T der Drehfeldmaschine 16 bei hoher Drehzahl Θ . ermittelt wird.The 2 shows the structure of a known Servo control according to 1 that around a device 42 is expanded to carry out the method according to the invention. The position, speed and current control loop 2 . 4 and 6 are a regulation 18 summarized, at the inputs of a rotor position actual value Θ, an actual speed value # ., a rotor position target value T *, a rotational speed target value Θ . * and the current components i _d and i _{q of} a determined motor current space vector

queue. In addition, this structure of the known servo control has a device 44 on, with the rotor position T of the induction machine in a known manner 16 at high speed Θ , is determined.

This device 44 has an EMF model 46 , An institution 48 for determining an error angle ε and a device 50 to determine the sign of the speed? , on. The determined error angle ε becomes a phase locked loop 52 supplied, which tracks the actual rotor position Θ in such a way that the error angle ε becomes zero.

an. Mittels diesem EMK-Modell 46 wird die ständerfeste Spannungsgleichung der Drehfeldmaschine 16 umgesetzt. Dieses EMK-Modell 46 ist auch als Spannungs-Modell bekannt. In den im Handel erhältlichen Stromrichtergeräten mit feldorientierter Regelung können anstelle eines Spannungs-Modells auch ein Strom-Modell bzw. ein Spannungs-Strom-Modell als EMK-Modell 46 verwendet werden.The EMK model 46 has a vector rotator on the voltage side 54 which converts the field-oriented voltage setpoint pointer u * / d, q into a stand-oriented voltage setpoint pointer u * / α,β with the help of the determined actual rotor position value Θ. On the current side, this EMF model 46 An institution 56 to determine current-dependent voltage drops, using a subtractor 58 from the stand-oriented voltage setpoint pointer u * / α,β be subtracted. At the exit of this subtractor 58 are the orthogonal components u _ΕΜΚα and u _{EMKβ of} the induced voltage space indicator

on. Using this EMF model 46 becomes the steady state voltage equation of the induction machine 16 implemented. This EMF model 46 is also known as a voltage model. In the commercially available converter devices with field-oriented control, instead of a voltage model, a current model or a voltage-current model can also be used as an EMF model 46 be used.

, der im kartesischen Koordinaten vorliegt, wird mittels eines Koordinatenwandlers 60 in polare Koordinaten umgewandelt, wobei die Winkelkoordinaten Θ_EMK mittels eines nachgeschalteten Vergleichers mit einem ermittelten Rotorlage-Istwert Θ verglichen wird. Ausgangsseitig steht dann ein Fehlerwinkel ε' an, der der Einrichtung 50 zur Bestimmung des Vorzeichens des Drehzahl-Istwertes Θ . zugeführt wird. Am Ausgang dieser Einrichtung 50 steht dann der korrekte Fehlerwinkel e an. Diese Einrichtung 50 weist eine Vorzeichenerkennung 64 und einen Vergleicher 66 auf. Mittels dieser Vorrichtung 44 in Verbindung mit dem Pha senregelkreis 52 kann der Rotorlage-Istwert T bei hoher Drehzahl Θ . ermittelt werden.The determined space pointer

, which is in Cartesian coordinates, is by means of a coordinate converter 60 converted into polar coordinates, the angular coordinates Θ _{EMF being} compared with a determined actual rotor position value Θ by means of a downstream comparator. An error angle ε 'then occurs on the output side, that of the device 50 to determine the sign of the actual speed value Θ , is fed. At the exit of this facility 50 the correct error angle e is then present. This facility 50 has a sign recognition 64 and a comparator 66 on. By means of this device 44 in connection with the phase control loop 52 can the actual rotor position value T at high speed? , be determined.

bestimmt werden.This known device 44 can only up to a lower speed Θ , _G which are operated, for example, 5% of the nominal speed Θ , _{n of} the induction machine 16 equivalent. Up to this lower limit speed Θ , _G can be an EMF space pointer

be determined.

mit einer nach geschalteten Additionseinrichtung 70 und andererseits eine Recheneinheit 72 mit nachgeschalteter Umrecheneinrichtunq 74 auf. Diese Additionseinrichtung 70 weist für jede feldorientierte Komponente d, q einen Addierer auf, so dass die Komponenten d, q des geforderten Spannungs-Sollraumzeigers

und die Komponenten d und q des generierten Testraumes

komponentenweise überlagert werden können. Dadurch entsteht ein Spannungs-Sollraumzeiger

der einen hochfrequenten Wechselanteil aufweist. Dieser Spannungs-Sollraumzeiger

wird mittels des Leistungsteils 12 an den Ständerwicklungen der Drehfeldmaschine 16 angelegt. Dadurch fließt in der Drehfeldmaschine 16 ein Motorstrom, der ebenfalls einen hochfrequenten Wechselanteil aufweist. Die orthogonalen Stromkomponenten i_dW und i_qW dieses Motorstrom-Raumanzeigers

werden ständerseitig ermittelt und mittels des Vektordrehers 38 in feldorientierte Stromkomponenten i_dW und i_qW transformiert. Mittels der Recheneinrichtung 72 wird das Testsignal vollständig vom Strom-Istraumzeigers

getrennt. Als Ergebnis erhält man an den beiden Ausgängen der Recheneinrichtung 72 ei nerseits einen testsignalbedingten Stromwechselanteil A_W und feldorientierte Komponenten d und q eines Strom-Istraumzeigers

, der keinen testsignalbedingten Stromwechselanteil A_W Around the rotor position Θ of a three-phase machine 16 even below this limit speed Θ , To be able to determine _G without having to use a mechanical sensor is the device 42 provided for carrying out the method according to the invention. This device 42 has a facility on the one hand 68 for generating a test space pointer

with an addition device connected downstream 70 and on the other hand an arithmetic unit 72 with subsequent conversion device 74 on. This adder 70 has an adder for each field-oriented component d, q, so that the components d, q of the required voltage target space vector

and the components d and q of the generated test room

can be overlaid component by component. This creates a voltage target space pointer

which has a high-frequency alternating component. This voltage target space pointer

is by means of the power section 12 on the stator windings of the induction machine 16 created. This causes the induction machine to flow 16 a motor current that also has a high-frequency alternating component. The orthogonal current components i _dW and i _{qW of} this motor current space indicator

are determined on the stand side and by means of the vector rotator 38 transformed into field-oriented current components i _dW and i _qW . By means of the computing device 72 the test signal is completely from the current Istra space pointer

Cut. The result is obtained at the two outputs of the computing device 72 On the one hand, a test signal-related current change component A _W and field-oriented components d and q of a current Istraum pointer

which has no test signal-related current change component A _W

has more. This test signal-related current change component A _W only depends on the test signal and the rotor position T of the induction machine 16 from. This conversion signal component A _W , which is proportional to the sine of the double error angle ε, is used to generate this test signal by means of the conversion device 74 the error angle e determines. In the steady state, this error angle e is zero, which means at an output of the phase locked loop 52 the true rotor position Θ of the induction machine 16 pending.

bestimmt werden. In der Umrecheneinrichtung 74 ist ebenfalls eine Rechenvorschrift hinterlegt. Diese lautet 0,5 × arcsin. Die Vorschriften der beiden Auswerteeinrichtungen 76 und 78 werden anhand der 6 und 7 näher erläutert.The computing device 72 has two evaluation devices 76 and 78 with which the determined current amplitudes a, b, c, d and e of the field-oriented current components i _dW and i _{qW are weighted} per switching period T _S and added to a current value. For this purpose, these evaluation devices 76 and 78 a rule with which the test signal-related current change component A _W and the current amplitude A of the field-oriented motor current Istraum pointer

be determined. In the conversion facility 74 a calculation rule is also stored. This is 0.5 × arcsin. The regulations of the two evaluation devices 76 and 78 are based on the 6 and 7 explained in more detail.

bezeichnet. Wenn eine Spannung an die ruhende Maschine 16 angelegt wird, dann ändert sich der Maschi nenstrom. Die dabei wirksame Induktivität ist in der d-Achse geringer als in der q-Achse, da der magnetische Fluss der Permanentmagnete 82 und 84 den Eisenkreis in der d-Richtung sättigt.In the 3 is the cross section through a induction machine 16 , in particular a servo motor. The rotor 80 this servo motor 16 has two permanent magnets 82 and 84 that generate a pathogen flow. The direction of this excitation flow is called the d-axis. If the rotor 80 turns in the positive direction, a voltage is induced perpendicular to it. This direction is called the q-axis. This induced voltage is also called the EMF voltage or EMF space vector

designated. When a voltage to the machine at rest 16 is applied, then the machine current changes. The effective inductance is less in the d-axis than in the q-axis because the magnetic flux of the permanent magnets 82 and 84 saturates the iron circle in the d direction.

When operating without a mechanical sensor, the true rotor position T is not known, so the position detection works with an encircling x, y coordinate system. In the 4 this rotating x, y coordinate system is shown with the fixed α, β coordinate system and with the rotating d, q rotor coordinate system. The d-axis of this d, q rotor coordinate system is in the direction of the excitation flux of the permanent magnet 82 and 84 generated excitation flow. This d, q coordinate system rotating with the rotor speed ω has a rotor position value Θ which is dependent on the rotor speed ω. In the 4 the circumferential x, y coordinate system is drawn behind the d, q coordinate system. The angle between these two rotating coordinate systems is e, which is called the error angle. In the method according to the invention, an estimated or determined initial position angle value is assumed. The result is an error angle e which indicates how far the x, y coordinate system is rotated relative to the d, q coordinate system. By means of the phase locked loop 52 the initial position angle value is tracked in such a way that the error angle e becomes zero. As soon as the error angle ε is zero, the guided initial position angle value corresponds to the actual rotor position value Θ of the induction machine 16 , In the 4 the x, y coordinate system and the d, q rotor coordinate system would then be congruent.

und

die jeweils aus zwei orthogonalen Komponenten bestehen. Zur Verdeutlichung des erfindungsgemäßen Verfahrens genügt, dass jeweils nur eine Komponente dieser Raumzeiger

und

dargestellt sind.In the 5 are several waveforms of the servo control according to the invention 2 each shown in a diagram over time t. The time span of these signal curves extends over three switching intervals n-1, n and n + 1. The signal profiles u * / d, q, u _T and u _W each show a component of the corresponding space pointer

and

each consisting of two orthogonal components. To clarify the method according to the invention, it is sufficient that only one component of these space pointers is used

and

are shown.

wird insbesondere so der Rotorlage Θ nahegeführt, dass der von der Testspannung verursachte Teststrom kein Drehmoment erzeugt. Diese reduziert deutlich die Geräuschentwicklung.The voltage component u * / d or u * / q is that of the current regulator 34 respectively. 32 ( 1 ) desired voltage component. The current regulator 32 respectively. 34 is carried out with a sampling interval T _S of 250 μsec at a switching frequency f _s = 4 kHz. The voltage values you want also change in this cycle. The actual values depend on the operation of the machine 16 off so that in this 5 shown waveforms are only to be understood as an example. In this example, the speed Θ , zero, whereas the current regulators 32 and 34 increase the current in the clock interval n. Limiting the current control cycle T _S to the switching frequency f _s does not limit the bandwidth of the speed control loop 4 because the sensible speed control bandwidth is anyway caused by the inaccuracy of the speed determined from the estimated position Θ , is limited. The signal curve u _T is the synchronous, rectangular test signal. The frequency and amplitude of this test signal can be independent of the operating state of the machine 16 be left constant. The direction of the vector

is particularly brought close to the rotor position insbesondere in such a way that the test current caused by the test voltage does not generate any torque. This significantly reduces noise.

The waveform u _W is the voltage that arises by adding u * / d or u * / q and u _T. This voltage is the pulse width modulator 10 supplied after it has been transformed into the stand-oriented α, β coordinate system. The signal curves u _r , u _s and u _t are those from the pulse width modulator 10 generated phase voltages. The signal _curve u _pwm is the voltage on the motor 16 (more precisely: a component of the vectorial voltage between the motor terminals). The pulse width modulator 10 ensures that the integral of this voltage corresponds to the desired voltage u _W over a modulation interval T _m of 125 μsec at a switching frequency f _s = 4 kHz. The signal _curve I _pwm is the resulting machine _current , in the example shown from a stationary machine 16 is assumed (i.e. no induced counter voltage). In reality, this stream is also a vector, but again only one component is shown here. It can be seen that the current is constant at the limits of the modulation intervals n-1, n and n + 1, since there the voltage u _{pwm is} zero. These values a _n , c _n and e _n for the sampling interval n indicated by a dot are required for the position detection. They can be determined either by a scanning current measurement at the interval limits or by an average current measurement around the interval limit. The dashed line shows the current curve, which would result if the current regulator 32 and 34 required voltage u * / q and u * / d on the machine 16 would be given. If you compare the actual current values a, c and e marked with dots with this dashed line, you can see that the values from a superposition of the desired one Current with an alternating component through the test signal.

The trapezoid below in the diagram of the 5 show the measuring intervals for the average current measurement. The mean values of the current over the intervals shown a . c and e correspond to the samples a, c and e if the zero pointer is at least as wide as the measuring interval. All of these values are vectors. Since the test signal and the test signal evaluation are only required at lower speeds, at which the induced voltage is also small, this is not a problem for the servo motor used. However, this results in a limit for the voltage and thus for the product of the test current amplitude and Frequency, so that the test signal frequency may have to be reduced if the motor inductance is high. This can be achieved either by reducing the switching frequency or by doubling the test signal period and the current control cycle at a constant switching frequency.

• – Das Testsignal soll mit der normalen Pulsbreitenmodulation realisierbar sein.
• – Das Testsignal sollte möglichst hochfrequent sein, da dann
• – eine schnellere Lageerkennung möglich ist
• – die Geräuschentwicklung reduziert wird.
• – Es soll gleichzeitig der Stromregelkreis betrieben werden.
• – Es soll eine gute Entkopplung der Lageerkennung von dynamischen Vorgängen möglich sein.
• – Der resultierende Strom muss mit der Stromistwerterfassung auswertbar sein.

In the practical implementation of the test signal, some boundary conditions have to be considered:

• - The test signal should be realizable with normal pulse width modulation.
• - The test signal should be as high-frequency as possible, because then
• - A faster position detection is possible
• - The noise is reduced.
• - The current control loop should be operated at the same time.
• - It should be possible to decouple the position detection from dynamic processes.
• - The resulting current must be evaluable with the current actual value acquisition.

The specified conditions can best be met with a test signal that is synchronous with pulse width modulation. The pulse width modulation with space vector modulation ensures that the voltage across the motor is averaged over a modulation period of the desired voltage u * / α, u * / β equivalent. The zero component is chosen such that: max (u r , u s , u t ) = –Min (u r , u s , u t )

Half a switching period (or a multiple thereof) can be selected as the modulation interval. With a modulation interval of half a switching period T _S , a test signal with a frequency equal to the switching frequency f _s can be generated.

In order to be able to operate the current control at the same time, the test signal is added to the voltage u * / d, q desired by the current controller. However, this results in the problem that dynamic control processes with correspondingly rapid voltage changes also contain high-frequency signal components, which can therefore lead to undesired interactions with the test signal u _T. This is avoided by the fact that the current controller may only output a new voltage value u * / d, q in every second modulation period T _M (for example by setting the sampling rate of the current control equal to the switching frequency f _s ).

To explain the determination of a test signal-related current change component A _W per switching period T _S as a function of determined current amplitudes a, b, c, d and e are in the 6 the voltage profiles u * / d, q, u _T and u _W with the associated current profiles i _{d, q} , i _T and i _W are shown in more detail. In the 7 in addition, the results of the test signal-related current change component A _W and the actual current values A are shown as a function of the two regulations for the test signal-dependent current change component A _W and the current actual value A. The regulation for determining the test signal-related current change component A _W is: A W = (2c - a - e) / 4 and the regulation for determining the actual current value A without the alternating current component A _W is: A = (0.5a + b + c + d + 0.5e) / 4.

Through these two regulations a complete Decoupling the test signal from the current control achieves what is illustrated by the numerical examples.

The voltage profile u * / d, q causes a current profile i _{d, q} in the machine 16 in the sampling interval n with the current amplitudes a = 0, b = 1, c = 2, d = 3 and e = 4. This results in a value for the alternating current component A _W = 0 and for the actual current value A = 2, depending on the specified regulations The test signal u _T causes a current profile i _T in the sampling interval n, the current amplitudes of which are as follows: a = −1, b = 0, c = 1, d = 0 and e = −1. This results in the following current values A _W = 1 and A = 0. The next diagram shows the voltage profile u _W resulting from the voltage profiles u * / d, q and u _T with the associated current profile i _W. In the sampling interval n, the current amplitudes are as follows: a = -1, b = 1, c = 3, d = 3 and e = 3. This results in a test signal-related current change component A _W = 1 and an actual current value A = 2. This shows that With the two specified regulations, a complete decoupling of the test signal, which, for example, generates a current change component A _W = 1 in the sampling interval n, is possible from the current control, which generates an actual current value A = 2 in the sampling interval n.

und der synchronen Strom-Auswertung erhält man eine sensorlose Lageerkennung im Stillstand und bei kleiner Drehzahl Θ ., die folgenden Vorteile:

• – hohe nutzbare Bandbreite
• – geringe Geräusche
• – Verwendung einer Standard-Pulserzeugung und eines Stan dard-Stromreglers,
• – vollständige Entkopplung von Stromregler und Testsignal

aufweist. Außerdem muss eine bestehende Servoregelung nicht neu konzipiert werden, sondern das erfindungsgemäße Verfahren zur Bestimmung einer Rotorlage Θ kann als separates Softwaremodul einfach in die bestehende Servoregelung integriert werden. Dieses Softwaremodul kann so integriert werden, dass es in Abhängigkeit vom Kundenwunsch aktiviert werden kann.With the field-oriented test room pointer synchronized with pulse width modulation

and the synchronous current evaluation, a sensorless position detection is obtained at standstill and at low speed Θ ., the following advantages:

• - high usable bandwidth
• - low noise
• - use of standard pulse generation and a standard current regulator,
• - Complete decoupling of current regulator and test signal

having. In addition, an existing servo control does not have to be redesigned, but the method according to the invention for determining a rotor position Θ can easily be integrated into the existing servo control as a separate software module. This software module can be integrated in such a way that it can be activated depending on customer requirements.

Claims (11)

Method for determining a rotor position (Θ) of a field-oriented induction machine ( 16 ), which has an effective inductance dependent on the rotor position (Θ), without a mechanical sensor at standstill and at low speed ((.) with the following process steps: a) provision of a test space pointer

, which is synchronous to the pulse width modulation of a field-oriented control of the field-oriented rotating field machine ( 16 ) is, b) component-wise superimposition of this test space pointer

for field-oriented components (u * / d, u * / q) of a setpoint voltage space vector required synchronously with pulse width modulation (

), c) Determination of at least two current amplitudes (a, ..., e) per switching period (T _S ) of a current space vector

the field-oriented induction machine ( 16 ), d) determining a test signal-related current change component (A _W ) per switching period (T _S ) as a function of the determined current amplitudes (a, c, e), e) converting this test signal-dependent current change component (A _W ) into an error angle (e), and f ) Tracking an estimated rotor position in such a way that the determined error angle (ε) becomes zero. A method according to claim 1, characterized in that the test space pointer (

) the rotor position (Θ) determined in such a way that it is in the d-axis of the field-oriented induction machine ( 16 ) lies. Method according to claim 1 or 2, characterized in that the test signal-related current change component (A _W ) according to the following equation: A W = (2c - a - e) / 4 is determined. Method according to one of the preceding claims, characterized in that the test space pointer (

) has a rectangular amplitude and its frequency is equal to the switching frequency (f _s ). Method according to one of the preceding claims, characterized in that, depending on the determined current amplitudes (a, ..., e), an actual current value (A) freed from the test signal-related current change component (A _W ) is determined. A method according to claim 5, characterized in that the actual current value (A) according to the following equation: A = (0.5a + b + c + d + 0.5e) / 4 is determined. Contraption ( 42 ) to carry out the method according to claim 1 with a field-oriented control ( 18 ) a field-oriented induction machine ( 16 ), characterized in that a device ( 68 ) to generate a test room pointer (

), an adder ( 70 ), a computing device ( 72 ) and a conversion facility ( 74 ) it is provided that the two outputs of the device ( 68 ) each with an adder of the adder ( 70 ) are connected, the second inputs of which are each connected to an output of a current control loop ( 32 . 34 ) of the field-oriented regulation ( 18 ) that the computing device ( 72 ) on the input side with outputs of a transformation device ( 38 ) and on the output side with inputs of the current control loop ( 6 ) and with an input of the conversion device ( 74 ) is linked and that the conversion device ( 74 ) on the output side with an input of a phase locked loop ( 52 ) connected is. Contraption ( 42 ) according to claim 7, characterized in that the device ( 42 ) is a software module. Contraption ( 42 ) according to claim 7 or 8, characterized in that the device ( 42 ) in the field-oriented regulation ( 18 ) is integrated. Contraption ( 42 ) according to one of claims 7 to 9, characterized in that the induction machine ( 16 ) with an effective inductance dependent on the rotor position (Θ) is a permanently excited synchronous machine. Contraption ( 42 ) according to claim 10, characterized in that the permanently excited synchronous machine is a permanently excited servo motor.