WO2019120380A1

WO2019120380A1 - Method and apparatus for controlling a dynamoelectric machine

Info

Publication number: WO2019120380A1
Application number: PCT/DE2018/101013
Authority: WO
Inventors: Arkady Pankevich
Original assignee: Schaeffler Technologies AG & Co. KG
Priority date: 2017-12-20
Filing date: 2018-12-13
Publication date: 2019-06-27
Also published as: DE102017130635A1

Abstract

The invention relates to a method and an apparatus for the field-oriented current control of a multiphase dynamoelectric machine (1) having a stator and a rotor which is at a distance therefrom via an air gap. In order to improve the dynamic and acoustic behaviour of the machine (1), the following method steps are proposed: i. transforming phase currents of the machine from a stator-based multiphase coordinate system (13) into a fundamental current vector in a two-axis first d/q coordinate system (14) rotating at a fundamental frequency, ii. transforming harmonic components of the phase currents of the machine from the stator-based multiphase coordinate system (13) into at least one harmonic current vector (16) in a two-axis second d/q coordinate system (15) which rotates at a harmonic frequency of the fundamental frequency, iii. determining a first manipulated variable vector, on the basis of a first system deviation between the fundamental current vector (5) and a fundamental desired current vector (2) in the first d/q coordinate system, which corresponds to a desired operating state of the machine, iv. determining a suitable harmonic desired current vector (10) for compensating for a harmonic component of the torque in the air gap of the dynamoelectric machine on the basis of the operating point of the dynamoelectric machine, v. determining a second manipulated variable vector on the basis of a second system deviation between the harmonic current vector (16) and the harmonic desired current vector (10), and vi. setting phase voltages for the electric machine (1) on the basis of the first and second manipulated variable vectors.

Description

Method and device for controlling a dynamoelectric machine

The invention relates to a method and a device for controlling a dynamo electric machine, in particular for controlling a permanent-magnet synchronous machine. The invention can i.a. are used for electric drive systems in an industrial environment as well as for electrically driven vehicles. In addition to all-electric motor vehicles and hybrid vehicles, an application for rail vehicles, for example, is also advantageous.

A widespread control concept of common 3-phase electrical machines, in particular permanent-magnet synchronous machines, is the so-called field-oriented control. This control concept is based on a transformation of the three-phase alternating variables into a two-axis coordinate system, which rotates synchronously with the rotor flux of the machine. In such a coordinate system, usually referred to as d / q coordinate system, for example, the three phase currents of the stator winding iu, iv, iw represented by a 2-dimensional current vector with the components i _q and id. In the case of an ideal sinusoidal rotor flux and ideal sinusoidal phase currents, the original alternating quantities iu, iv, iw are mapped to DCs i _q , id as a consequence of the coordinate system rotating synchronously with the rotor flux. In such a coordinate system, the q-component of the current represents the torque-forming component of the machine analogously to the armature current of a DC machine. The d-component of the current represents the field-forming component of the machine current and corresponds to the exciter current of a DC machine. Thus, the transformation achieves that the three-phase machine can be controlled in a similar advantageous manner as a DC machine, in which the active and reactive power components can be controlled independently of one another.

However, the field-oriented control experiences its limits when the air gap flow of the machine deviates greatly from the sinusoidal shape. Such deviations are due to machine geometry and especially in the widespread tooth coil wound usually very pronounced. With these windings, the pronounced teeth of the stator, viewed over the circumference of the air gap, lead to different magnetic conductivities. As a result, harmonics occur in the air gap flux which, in turn, results in reluctance-related cogging moments (so-called cogging). The harmonic components in the air gap flow are undesirable because they usually do not contribute to the torque of the machine but merely increase the iron losses. In addition, cogging adversely affects the acoustic properties of the machine.

The flux harmonics induce voltages in the stator windings which, in turn, result in a harmonic content in the phase currents. This results in harmonic components in the phase current with the 5, 7, 13, 17 ... -fold frequency of the fundamental. The above-mentioned current harmonics can as a rule only be eliminated in a very limited frequency range by the methods known from the prior art and based on the field-oriented regulation. From "Robert Michel, compensation of saturation-induced harmonics in the currents of field-oriented synchronous motors, Dissertation Technical University Dresden, 2009 / Chapter 4.2" Although a method is known with which the harmonic components of the q-current and d-current at a machine control based on the field-oriented control. However, this is based on a feedforward control, which requires a complex investigation and modeling of the electrical machine to be controlled. In the case of changed environmental conditions, machine parameters may deviate that deviate from the previously created model so that feedforward control no longer leads to the desired result.

Furthermore, a compensation of the harmonic content in the phase currents leads to lower copper losses in the windings. However, it does not necessarily result in improved acoustic properties of the machine. To improve these, cogging must be avoided. This goal can be achieved by specifically imparting harmonic components to the phase current, which result in a reduction of the harmonic components in the magnetic air gap moment. The amplitude and phase of these harmonic components depend on the operating point of the electrical machine and its properties. From US Pat. No. 8,541,968 B2, a method for determining suitable compensation signals is known in which the latter are determined operating point-dependent by means of two test runs. In this case, two different compensation signals are applied to the machine and the corresponding vibration responses of the machine are determined. A mathematical operation can be used to calculate the ideal compensation signal on the basis of these two test runs.

The object of the invention is to improve the dynamic and acoustic behavior of an electrical machine, in particular a synchronous machine, in as little effort as possible.

This object is achieved by a method having the features according to patent claim 1 and by a device having the features according to patent claim 8.

Advantageous embodiments of the invention can be found in the dependent patent claims.

The core idea of the invention is based on the finding that the dynamics of the control circuit for the phase currents, in particular for harmonic phase currents, of the machine can be significantly improved if separate control circuits are provided for the fundamental and the harmonic (n), respectively individually parameterized. The harmonics are understood to be the integer multiples of the fundamental. The improved dynamics of the current regulation ultimately improve the acoustic properties of the electric drive train.

In order to control the active and reactive power components of the machine current, the phase currents are transformed from a stator-related polyphase, in particular 3-phase, coordinate system into a fundamental current vector in a biaxial first d / q coordinate system rotating at a fundamental frequency. In this method step, the well-known Clarke transformation is performed first. As an example of a 3-phase machine, this means that the three Phase currents of the machine are transferred from the stator fixed 3-phase coordinate system in a likewise fixed to the stator 2-phase coordinate system. Following this, the already two-dimensional current vector is converted by means of the so-called park transformation, also d / q transformation, into a coordinate system rotating synchronously with the rotor flux. The direct current component of the d-current id corresponds to the exciting part of the current attributable to the fundamental of the current, while the dc component of the q-current represents the torque-forming fundamental component of the stator current.

If, for example, the torque demand on the electric machine is increased, a higher-level control will prescribe an altered setpoint for the q component of the current in the first d / q coordinate system, which is to be set by the current controller with the highest possible dynamics and stability , According to the invention, a first manipulated variable vector is determined on the basis of a first control deviation between the fundamental oscillation current vector and a fundamental oscillation current vector, the latter undergoing a sudden change in this example due to the changed torque specification. Also, the fundamental harmonic current vector is in the form of d / q coordinates and is related to the first d / q coordinate system. A controller provided for determining the manipulated variable vector on the basis of the system deviation can be dimensioned and parameterized for the expected fundamental frequency range of the machine. Since this controller does not at the same time have a highly dynamic imprinting of harmonic components to compensate for harmonic components in the torque of the machine, they do not have to be taken into account in the parameterization of the controller either.

In a further method step according to the invention, harmonic components of the phase currents of the machine from the stator-related multiphase coordinate system are transformed into at least one harmonic current vector in a second biaxial d / q coordinate system which rotates at a harmonic frequency of the fundamental frequency. Again, the well-known Clarke and Park transformations are performed. Before the transformations are carried out, the harmonic components can first be removed from the Phase currents of the machine are extracted. However, this is not absolutely necessary. It is also conceivable and encompassed by the invention to transform the complete phase currents, ie with fundamental and harmonic components, in this process step.

Based on this, in a further method step, a second manipulated variable vector is determined on the basis of a second control deviation between the harmonic current vector and a harmonic desired current vector.

The harmonic setpoint current vector is unequal to the zero vector in a real electrical machine in which a rotation angle-dependent ripple moment sets. Because an ideal sinusoidal phase current in this case has no cogging freedom result. The air gap moment is a function of the q component of the current and the magnetic flux. Thus, harmonics in the air gap moment can also occur as a result of flux harmonics, even if the phase windings of the stator are fed with an ideally sinusoidal current. Such flux surges are due to the geometry of the electrical machine.

The aim of the current control according to the invention is therefore to compensate for the ripple of the air gap moment by impressing suitable harmonic components in the phase current. The harmonic setpoint current vector thus denotes a quantity that is suitable for compensating a harmonic component of the moment in the air gap of the electrical machine. He thus designates a component of electricity whose primary purpose is the suppression of cogging with respect to a harmonic. The components of the harmonics nominal current vector suitable for this purpose are dependent on the operating point of the electrical machine, i. H. in particular of their torque and speed. In addition, the geometric and magnetic properties of the machine are decisive for the compensation currents with which cogging can be effectively suppressed.

Operating point-dependent harmonic setpoint current vectors, which can solve the compensation task described above, can be determined, for example, by simulating the machine or the complete drive train in which the machine is embedded. Alternatively, the operating point-dependent upper harmonic setpoint current vectors can be determined experimentally, as proposed, for example, in the already mentioned US Pat. No. 8,541,968 B2. The determined values are stored, for example, subsequently in the form of a look-up table in a memory, to which a controller has access for determining suitable switching signals for a converter. By contrast, the form of the data for the harmonic setpoint current vectors is arbitrary, as long as it is ensured that they can be transformed to a form which can be represented in the second d / q coordinate system. Thus, for example, the harmonic desired current vectors can be directly stored as two-dimensional vectors with the variables id and _iq , which are related to the second d / q coordinate system. Alternatively, however, a representation is conceivable in which the harmonic setpoint current vectors are stored in the form of amplitude and phase angle in the look-up table.

A controller provided for the determination of the second manipulated variable vector can be optimized with regard to its parameterization specifically to the expected frequency range of the second manipulated variable vector. The regulation of the fundamental component is just as little his task as the control of other harmonics that do not correspond to its harmonic frequency. It is also possible to provide a filter which extracts the DC component from the harmonic current vector so that only the specific harmonic which is the controlled variable is used as the input variable for the controller.

This transformation and the subsequent control of the associated harmonic current vector are carried out according to the invention for at least one harmonic of the phase currents; in an advantageous embodiment of the invention, however, for several harmonics, in particular for all harmonics that have a sig- nificant impact on the acoustics and the losses of the machine. In this case, harmonics nominal current vectors, depending on the operating point of the electrical machine, can be stored in the look-up table for the different harmonics and read out for current regulation.

Finally, in a further method step, phase voltages for the electrical machine are dependent on the first and second manipulated variable vector. posed. In the case of a transformation and regulation of further harmonic components, the further manipulated variable vectors derived therefrom are also correspondingly taken into account in the adjustment of the phase voltages.

The control method according to the invention has the advantage over the prior art that targeted, highly dynamic control of the harmonic component is possible without determining and modeling in detail the magnetic and mechanical structure of the dynamoelectric machine. The latter is the case with known methods which are based on control of the harmonic components by disturbance step-up, which makes these methods impractical for many applications.

In an advantageous embodiment of the invention, the first manipulated variable vector represents the output variable of a first controller and the second manipulated variable vector represents the output variable of a second controller, which differs from the first controller. By the use of a separate controller for the fundamental component and In each case, one controller for each individual harmonic component to be controlled can be used to implement a current regulation with higher disturbance dynamics. The dimensioning of each individual controller over a narrow frequency range enables greater dynamics than can be achieved by a single current regulator that has to operate over a wide frequency range.

A further advantageous embodiment of the invention is characterized in that the first and second controllers are differently parameterized PI controllers. Alternatively, a PID controller for the first and second controller can be provided.

In a further advantageous embodiment of the invention, all manipulated variable vectors are transformed by means of inverse Clarke transformation into a stator-related 2-phase a / b coordinate system and then summed into a manipulated variable sum vector. On the basis of the manipulated variable sum vector, switching signals for a converter feeding the machine can then be generated by means of a space vector modulation.

An apparatus according to the invention for field-oriented current regulation of a multi-phase dynamoelectric machine comprises a processing unit for the a. Transforming phase currents of the machine from a stator-related polyphase coordinate system into a fundamental current vector in a biaxial first d / q coordinate system rotating at a fundamental frequency;

b. Transforming harmonic components of the phase currents of the machine from the stator-related polyphase coordinate system into at least one harmonic current vector in a biaxial second d / q coordinate system that rotates at a harmonic frequency of the fundamental frequency;

c. Determining a first manipulated variable vector on the basis of a first control deviation between the fundamental oscillation current vector and a basic oscillation nominal current vector in the first d / q coordinate system,

d. Determining a suitable harmonic setpoint vector for compensating a harmonic component of the magnetic moment in the air gap of the dynamoelectric machine as a function of the operating point of the dynamoelectric machine, and

e. Determining a second manipulated variable vector on the basis of a second control deviation between the harmonic current vector and the harmonic desired current vector,

wherein the apparatus further comprises a memory for a look-up table, in which harmonic desired current vectors for different operating points of the machine are stored, and wherein the device comprises a converter for adjusting phase voltages for the electrical machine in dependence on the first and second manipulated variable vector.

An electric drive system with a permanently excited synchronous machine and such a device is characterized by lower losses and better acoustic properties in comparison to known electric drive systems.

In the following, the invention will be explained in more detail with reference to the embodiments illustrated in the figures. Functionally identical elements are provided in the figures with the same reference numerals.

Show it: FIG. 1 shows a control circuit known from the prior art for a synchronous machine on the basis of the field-oriented control;

FIG. 2: a stator-fixed 3-phase coordinate system and a d / q coordinate system rotating with the fundamental oscillation of the rotor flux according to the prior art,

3 shows a control circuit for a synchronous machine according to an embodiment of the invention and

FIG. 4 shows a 3-phase stator-fixed coordinate system and d / q coordinate systems rotating synchronously with harmonics of the rotor flux according to an embodiment of the invention.

FIG. 1 shows a control circuit known from the prior art for a permanent magnet synchronous machine 1, as is well known from the prior art. For example, the permanent-magnet synchronous machine 1 is a drive motor of an electric vehicle. From an actuation of the accelerator pedal and the speed of the vehicle, a higher-level controller determines desired values for the torque and the rotational speed of the synchronous machine 1. A higher-order controller uses these variables to determine a fundamental desired torque vector 2 for the stator winding in the stator winding Synchronous machine 1 feed the current. This fundamental desired flux vector 2 is related to a rotor flux-fixed d / q coordinate system and therefore comprises a component i _q ^* which is proportional to the torque of the synchronous machine 1 and a component id ^* which influences the air gap flux within the machine. The fundamental oscillation setpoint vector 2 forms the reference variable of the control loop.

The synchronous machine 1 comprises a rotor position sensor 3 for determining the rotor angle Q. The rotor position is required to carry out a rotor flux-oriented d / q transformation 4, which is known by the name Clarke Parks transformation. Input variables for the d / q transformation 4 are, in addition to the rotor position angle Q, the three phase currents i _u , iv, iw. Through the d / q transformation 4 these become three phase currents are transformed into a fundamental oscillation current vector 5 with the components i _q , id.

Finally, the control deviation between the fundamental nominal flux vector 2 and the fundamental oscillation current vector 5, which is used as input for a PI controller 6, is determined. Finally, at the output of the PI controller 6 there is a manipulated variable vector in the d / q coordinate system, which is converted by means of an inverse Clarke transformation 7 into a still 2-dimensional, but now stator-fixed, coordinate system. The manipulated variable vector in the so-called a / ß coordinate system is outstandingly suitable for generating a corresponding switching pattern for a three-phase voltage source converter 9 by means of a so-called space vector modulation 8. The voltage source converter 9 is finally connected to the terminals of the stator winding of the synchronous machine 1 connected.

FIG. 2 shows the coordinate systems before the d / q transformation with the u, v, w coordinates and after the d / q transformation with the coordinates q, d. The d / q coordinate system is at the rotor flux of the machine so that the fundamental components of the current in the d / q coordinate system are mapped as equal. However, harmonic components of the phase current are also alternating variables in the d / q coordinate system. If one now wants to simultaneously achieve a highly dynamic and stable control of the fundamental component and a targeted adjustment of harmonic components in the phase current in order to compensate for harmonic components of the air gap torque, the problem arises of finding suitable parameters for a single controller, which comparatively large frequency range must serve.

Figure 3 shows an embodiment according to the invention, which solves the problem addressed. FIG. 4 shows the associated coordinate systems.

As can be seen in FIG. 3, the control circuit contains, in addition to a first control variable in the form of the fundamental desired torque vector 2 already shown in FIG. 1, a harmonic desired current vector 10 with the components i _q s ^* and ids ^* . This Harmonic desired current vector 10 represents a setpoint specification for the fifth harmonic of the phase current. The value of this harmonic desired current vector has been read from a look-up table, which is stored in a memory, not shown. For various operating points of the electric machine, ie for different torques and speeds, this look-up table contains suitable compensation signals in the form of harmonic setpoint current vectors. These have been previously determined by means of a finite element simulation or experimentally by determining the vibration responses of the machine or drive train in which the machine is embedded for different harmonic components in the phase current. Accordingly, the illustrated control circuit triggers the task of regulating the fifth harmonic of the phase current to the harmonic desired current vector, which was read from the look-up table as suitable compensation signals for the operating point aimed at by the machine.

It should be noted at this point that the harmonic setpoint harmonic vector for the fifth harmonic 10 is merely one example of an additional control variable. Expediently, in each case a functional relationship between an operating point of the machine and an associated harmonic desired current vector is stored in the look-up table for several, in particular all harmonic components that significantly occur in the magnetic air gap field. Each of these harmonic setpoint current vectors then represents its own control variable for its own control loop. For the sake of clarity, the representation in FIG. 3 contains only one control circuit for the fifth harmonic.

In principle, the control circuit in FIG. 3 corresponds structurally to the control circuit in FIG. 1. However, an additional d / q transformation 11 is provided for transforming the fifth harmonic of the phase currents into the d / q system and an additional inverse Clarke transformation 12 for back transformation from the d / q system to the a / b system. The output of the additional block for the d / q transform is finally subtracted from the fifth harmonic setpoints to determine the fifth harmonic error. This control deviation is then transferred by further PI controller 17 in a manipulated variable vector. Subsequently, by means of the additional Clarke transformation 12, a ne representation of the manipulated variable vector in the a / b system generated. Before the generation of the switching nut pattern by means of the space vector modulation 8, the manipulated variable vectors for the fundamental oscillation component and the harmonic component are added. As shown in Figure 4, the gist of the invention is to perform the d / q transformation for the fundamental fraction as known in the art. This means that the electrical variables originally defined in a stator-fixed 3-phase coordinate system 13 are converted into a biaxial coordinate system 14 rotating at a fundamental frequency. In addition, however, each relevant harmonic component is transformed into another two-axis d / q coordinate system 15, which rotates with a corresponding frequency of the harmonic component. There, the harmonic component is regulated to its nominal value with its own controller, which is optimally parameterized with regard to the frequency range of the harmonic.

Reference character list synchronous machine

Fundamental target current vector

Rotor position sensor

d / q transformation

Fundamental current vector

first PI controller

Inverse Clarke transformation

Space vector modulation

voltage source

Harmonic target current vector

additional d / q transformation

additional inverse Clarke transformation

stator-related 3-phase coordinate system

a two-axis coordinate system rotating at a fundamental frequency

another biaxial d / q coordinate system

Harmonic current vector

second PI controller

Claims

claims

1. Method for field-oriented current regulation of a multiphase dynamoelectric machine (1) having a stator and a rotor spaced apart therefrom via an air gap, with the following method steps:

i. Transformation of phase currents of the machine from a stator-related multiphase coordinate system (13) into a fundamental-wave current vector in a biaxial first d / q-coordinate system (14) rotating at a fundamental frequency;

ii. Transforming harmonic components of the phase currents of the machine from the stator-related polyphase coordinate system (13) into at least one harmonic current vector (16) in a biaxial second d / q coordinate system (15) rotating at a harmonic frequency of the fundamental frequency;

iii. Determining a first manipulated variable vector on the basis of a first control deviation between the fundamental oscillation current vector (5) and a basic oscillation nominal current vector (2) in the first d / q coordinate system that corresponds to a desired operating state of the machine,

iv. Determination of a suitable harmonic setpoint current vector (10) for

Compensation of a harmonic component of the moment in the air gap of the dynamoelectric machine as a function of the operating point of the dynamoelectric machine,

v. Determining a second manipulated variable vector on the basis of a second control deviation between the harmonic current vector (16) and the harmonics nominal current vector (10) and

vi. Setting phase voltages for the electric machine (1) as a function of the first and second manipulated variable vector.

2. The method of claim 1, wherein the first manipulated variable vector represents the output of a first controller (6) and the second manipulated variable vector represents the output of a second controller (17), which is different from the first controller (6).

3. The method of claim 2, wherein the first and second controllers (6, 17) are differently parameterized PI controllers.

4. Method according to one of the preceding claims, wherein according to the method steps ii, iv and v according to claim 1 further manipulated variable vectors are determined on the basis of a control deviation between further harmonic current vectors (16) and harmonics nominal current vectors (10) and in the setting of the pha - Sens voltage according to step v be considered.

5. A method according to any one of the preceding claims, wherein the one or more harmonic setpoint current vectors (10) which are suitable for compensating for a harmonic component of the moment in the air gap of the dynamoelectric machine in dependence on the operating point of the dynamoelectric machine, from a look-up table be read out.

6. Method according to one of the preceding claims, wherein all manipulated variable vectors are transformed by means of inverse Clarke transformation (7, 12) into a stator-related 2-phase a / b coordinate system and then summed to form a manipulated variable sum vector.

7. The method of claim 6, wherein on the basis of the manipulated variable sum vector with the aid of a space vector modulation (8) switching signals for a machine (1) speci send inverter (9) are generated.

8. Device for field-oriented current control of a polyphase dynamoelectric machine (1), which comprises a processing unit for

a. Transforming phase currents of the machine from a stator-related polyphase coordinate system (13) into a fundamental current vector (5) in a biaxial first d / q coordinate system (14) rotating at a fundamental frequency;

b. Transforming harmonic components of the phase currents of the machine from the stator-related polyphase coordinate system (13) into at least one harmonic current vector (16) in a biaxial second d / q coordinate system (15) rotating at a harmonic frequency of the fundamental frequency; c. Determining a first manipulated variable vector on the basis of a first control deviation between the fundamental oscillation current vector (5) and a basic oscillation nominal current vector (2) in the first d / q coordinate system,

d. Determining a suitable harmonics nominal current vector (10) for compensating a harmonic component of the moment in the air gap of the dynamoelectric machine as a function of the operating point of the dynamoelectric machine, and

e. Determining a second manipulated variable vector on the basis of a second control deviation between the harmonic current vector (16) and the harmonics nominal current vector (10),

wherein the device further comprises a memory for a look-up table, in which harmonic setpoint current vectors (10) for different operating points of the machine are stored, and wherein the device comprises a converter (9) for setting phase voltages for the electrical machine (1) in dependence on the first and second manipulated variable vector.

9. An electric drive system with a permanent-magnet synchronous machine and a device according to claim 8.