WO2021198264A1

WO2021198264A1 - Method for capturing a rotor position of a permanently excited synchronous machine, and permanently excited synchhronous machine having means for capturing the rotor position thereof

Info

Publication number: WO2021198264A1
Application number: PCT/EP2021/058310
Authority: WO
Inventors: Martin Grotjahn; Nils WILCKEN; Daniel Engelke; Thomas Eutebach
Original assignee: Hochschule Hannover
Priority date: 2020-04-01
Filing date: 2021-03-30
Publication date: 2021-10-07
Also published as: DE102020109004A1

Abstract

In order to capture a rotor position of a rotor (2) of a permanently excited synchronous machine (1), voltages (27) are applied to stator windings of the synchronous machine (1), wherein higher-frequency additional signals (16) are superimposed on fundamental components (14) of the voltages (27), which are applied for the purpose of operating the synchronous machine (1), at least at low speeds of the synchronous machine (1). As responses to the applied voltages, currents (19-21) flowing through the stator windings are measured. An observer for estimating the rotor position (22) of the permanently excited synchronous machine (1) on the basis of the voltages applied to the stator windings and the currents flowing through the stator windings is defined as an extended Kalman filter (EKF) (12) which comprises a model (13) of the synchronous machine (1). The model (13) of the synchronous machine (1) has a fundamental model component and an additional signal model component. The fundamental components (14) and the additional signals (16) of the applied voltages (27) and the measured currents (19-21) are continuously input to the EKF (12). The EKF (12) is used as the sole observer for the rotor position (22) at all speeds of the synchronous machine (1), with the result that two or more different observers for estimating the rotor position (22) of the permanently excited synchronous machine (1) do not need to be synchronized on the basis of the speed.

Description

METHOD OF DETECTING A ROTOR POSITION OF A PERMANENTLY EXCITED SYNCHRONOUS MACHINE AND PERMANENTLY EXCITED SYNCHRONOUS MACHINE WITH MEANS

TO DETERMINE YOUR ROTOR POSITION

TECHNICAL FIELD OF THE INVENTION

The invention relates to a method for detecting a rotor position of a rotor of a permanently excited synchronous machine with the features of the preamble of independent claim 1 and to a permanently excited synchronous machine with the features of the preamble of independent claim 8. In particular, the permanently excited synchronous machine (PSM ) to a sensorless permanent magnet synchronous machine (SLPSM).

STATE OF THE ART

Sensorless permanent magnet synchronous machines (SLPSM) represent an alternative to asynchronous machines that combine high energy efficiency and low costs. With SLPSM, however, standstill and very low speeds are difficult to control because known methods for detecting the rotor position are too imprecise or, due to their design, are not well suited for many SLPSMs. This is why the SLPSM often uses a controlled start-up or pole latching method at low speeds. However, these are energetically unfavorable and result in a significantly poorer energy efficiency when a SLPSM is operated frequently at low speeds, as occurs for example with conveyor belts with changing speeds or fans with varying speeds.

In principle, the rotor position of an SLPSM can be reconstructed from measured currents that flow through the stator windings of the SLPSM as responses to applied voltages. There the currents through the stator windings have to be measured anyway for the regulation of a SLPSM, no additional sensors are required for this.

Many known methods for sensorless control of permanent magnet synchronous machines can be divided into anisotropy and EMF-based methods. In anisotropy-based methods, a difference in inductances in the d- and q-direction of the synchronous machine generates an angle-dependent sequence in the currents that flow through the stator windings in response to the applied voltages. By superimposing the fundamental wave components of the voltages that are applied to operate the synchronous machine with higher-frequency additional signals, the information content regarding the rotor position in the currents flowing in response to the applied voltages can be increased. The information about the rotor position can be isolated from the currents flowing in response to the applied voltages by means of various systems. In the field of position sensorless control, these systems are designed, for example, as so-called angle trackers. A well-known and frequently used angle tracker uses a phase locked loop (PLL). A major disadvantage of the anisotropy-based method is the high dependency on the anisotropy of the inductances and thus on the structural properties of the synchronous machine. In addition, depending on the synchronous machine, the anisotropy depends to a different extent on the saturation state of the synchronous machine, so an anisotropy-based method can be unsuitable at certain operating points of the respective synchronous machine. In addition, the additional current components flowing through the stator windings as responses to the higher-frequency additional signals have an effect on the superimposed control of the synchronous machine. Filters are therefore used to isolate the fundamental wave current components from the measured currents, for example band-stop filters or circulating low-pass filters. However, these filters reduce the signal quality of the fundamental current components and generate an additional phase delay. This affects the regulation of the synchronous machine. Another disadvantage is with the angle trackers, which work, for example, according to the PLL method, because they have a transient response and thus delay the information about the rotor position. Known anisotropy-based methods in which the rotor position is calculated directly and which therefore have no transient behavior, on the other hand, have other disadvantages such as insufficient accuracy and very high susceptibility to failure. EMF-based methods use the information about the rotor position, which is contained in the electromotive force (EMF). Using linear observers, it is possible to estimate the EMF and to determine the rotor position using an angle tracker. Based on a model, these observers estimate the currents flowing through the stator windings of the synchronous machine, which are then compared with the measured currents. This results in the observer error, which is used to adapt the model. Non-linear observers such as the Extended Kalman Filter (EKF) are an extension of the position sensorless control of permanent magnet synchronous machines. Non-linear observers can also map a non-linear system behavior in a model-based manner and can thus be used to estimate the rotor position without the need for an additional angle racker. Depending on the quality of the underlying model, it is possible to significantly reduce the phase delay of the estimated rotor position compared to conventional angle trackers. However, the non-linear observers also obtain their information about the rotor position from the EMF, which is directly dependent on the speed, in particular proportional to the speed and is therefore zero when the synchronous machine is at a standstill. EMF-based methods can therefore only be used above a certain minimum speed of the synchronous machine.

Because of the specific advantages and disadvantages of anisotropy and EMF-based methods, it is not uncommon to combine two observers for the rotor position in order to use the respective advantages and to compensate for the disadvantages. With such a combination, a switch is made between the two methods for estimating the rotor position as a function of the speed.

There are various options for switching, such as hard or ramp-shaped switching between the two observers. The switchover itself, but also the transient response of the individual observers, has a negative effect on the superimposed control of the synchronous machine. From CN 108599661 A it is known to estimate the speed and rotor position of the rotor of a permanent magnet synchronous machine at medium and high speeds using an observer defined as an Extended Kalman Filter (EKF). At low speed, the speed and the rotor position are estimated using an anisotropy-based method, in which the fundamental wave components of the voltages applied to the synchronous machine's stator windings to operate are superimposed with higher-frequency additional signals. That

Switching between the two observers for the speed and the rotor position takes place above a minimum speed at which the EKF works stably. From M. Bugsch and B. Piepenbreier: "High-Bandwidth Sensorless Control of Synchronous Reluctance Machines in the Low- and Zero-Speed Range", IEEE Transactions on Industry Applications, Vol. 56, No.3, pages 2663 to 2672 is one sensorless control of synchronous reluctance machines in the range of low speeds up to standstill known. The control is intended to limit a current ripple, as it is caused by an excitation voltage in the form of a rectangular function, and at the same time limit the amplitude of a required HF test current. The control proposed for this purpose is optimized for the highly non-linear machine behavior of synchronous reluctance machines. The control includes low-noise current demodulation through a Kalman filter with empirical tuning. Specifically, a machine parameter-free Kalman filter is proposed. The quantities estimated with the Kalman filter are the HF current ripple and the fundamental wave current component. These variables are estimated based on the excitation voltages for the fundamental wave current component and the HF test current as well as measured values of the currents actually flowing. A control of the speed and the fundamental wave components of the synchronous reluctance machine, which is otherwise independent of the Kalman filter, is based on the estimated currents.

OBJECT OF THE INVENTION

The invention is based on the object of providing a method for detecting a rotor position of a rotor of a permanent-magnet synchronous machine and a permanent-magnet synchronous machine which precisely determine the rotor position at different speeds of the permanent-magnet synchronous machine. The method should be robust and suitable for a large number of permanent magnet synchronous machines.

SOLUTION

The object of the invention is achieved by a method for detecting a rotor position of a permanently excited synchronous machine with the features of independent claim 1 and by a permanently excited synchronous machine with the features of the independent

Claim 8 solved. The dependent claims relate to preferred embodiments of the method according to the invention and the permanently excited synchronous machine according to the invention. Claim 15 is directed to a permanent magnet synchronous machine with devices for carrying out the method according to the invention. DESCRIPTION OF THE INVENTION

In the method according to the invention for detecting a rotor position of a rotor of a permanent magnet synchronous machine, voltages are applied to the stator windings of the synchronous machine, with higher-frequency additional signals superimposed on the fundamental wave components of the voltages applied to operate the synchronous machine at low speeds of the synchronous machine. Currents through the stator windings flowing in response to the applied voltages are measured. An observer for estimating the rotor position of the permanent magnet synchronous machine based on the voltages applied to the stator windings and the currents flowing through the stator windings is defined as an Extended Kalman Filter (EKF). The EKF contains a model of the synchronous machine which, according to the invention, has a fundamental wave model component and an additional signal model component as parts of its closed mathematical representation in the state space. The fundamental wave components and the additional signals of the applied voltages and the measured currents are used as input signals to the EKF and are continuously entered into the EKF. The EKF is used as the only observer for the rotor position at all speeds of the synchronous machine, i. That is, there is no speed-dependent switchover or other synchronization between two or more different observers for estimating the rotor position of the permanent-magnet synchronous machine.

The fact that in the method according to the invention the fundamental wave components of the voltages applied to operate the synchronous machine are superimposed with higher-frequency additional signals at low speeds of the synchronous machine does not expressly exclude that this superimposition also takes place at higher speeds. In other words, the additional signals are superimposed at least at the low speeds, in particular in a speed range starting at zero. In order to use the fundamental wave components and the additional signals of the applied voltages as input variables of the EKF, measured values of the fundamental wave components and the additional signals or corresponding target values for these components of the applied voltages can be used. For example, the voltages applied to the stator windings of the synchronous machine on the one hand and their fundamental wave components or the additional signals on the other hand can be continuously entered into the EKF because their difference provides the missing components, ie the additional signals or the fundamental wave components. The model of the synchronous machine can be a purely electrical model. However, a mechanical model component can also be provided as part of the closed mathematical representation of the model of the synchronous machine in the state space. By integrating the mechanical part of the model of the synchronous machine in the equations of state of the EKF, the accuracy and robustness with which the EKF determines the rotor position in operating points with little information content of the measurements available, such as B. at very low speeds of the synchronous machine or when the machine is very busy.

The EKF can also be used to estimate the fundamental current components of the currents that flow through the stator windings as responses to the fundamental wave components of the applied voltages and to evaluate them with regard to the rotor position of the synchronous machine. The estimated fundamental current components can then be used together with the estimated rotor position in order to regulate the fundamental current components to desired values. Since the fundamental wave current components of the measured currents are estimated by the EKF and not isolated from the measured currents, for example by bandstop filters or with rotating low-pass filters, neither the signal quality of the fundamental wave current components is reduced nor a phase delay is introduced. This provides an optimal basis for regulating the fundamental wave current components for optimal operation of the synchronous machine.

The EKF can also estimate additional current components of the currents that flow through the stator windings as responses to the higher-frequency additional signals of the applied voltages and also evaluate them with regard to the rotor position of the synchronous machine.

In addition, the EKF can estimate an angular speed of the rotor as a further relevant system state of the synchronous machine in addition to the rotor position, the fundamental wave current components and the additional current components. In particular, if the model of the synchronous machine has a mechanical model component in its additional signal model component, the EKF can make this estimate directly, i. H. without going through a differentiation of the rotor position.

In the method according to the invention, an additional signal specification for the additional signals in the form of a rotating voltage vector can be superimposed on a voltage manipulated variable for the fundamental wave components of the applied voltages. More specifically, the additional signals in the alpha / beta coordinate system can be a cosine signal in the alpha direction and a sine signal of the same kind Include amplitude in the beta direction and thus result in a voltage vector that circles around the origin.

A frequency of the additional signals can be at least twice as great as a rated frequency of the synchronous machine. There is no hard upper limit for the frequency of the additional signals. Typically, however, a voltage regulator for the voltages applied to the stator windings of a synchronous machine is only capable of realizing additional signals of limited frequency. As a rule, therefore, in the method according to the invention, the frequency of the additional signals will not exceed 100 times the rated frequency of the synchronous machine. The rated frequency of a synchronous machine is the frequency of the fundamental wave components of the voltages applied to the stator windings of the synchronous machine, for which the synchronous machine and its voltage regulator are designed.

In the method according to the invention, the EKF can also estimate further state variables and system parameters of the synchronous machine. The EKF can estimate a voltage-inducing portion of the flux linkage of permanent magnets on the rotor of the synchronous machine and / or a mechanical disturbance variable in the stator windings. This mechanical disturbance variable can in particular estimate external influences on the rotor, such as those that occur when the synchronous machine is used as a drive for a conveyor belt, for example. In a permanently excited synchronous machine according to the invention with a rotor, with stator windings, with a voltage regulator, which is designed and arranged to apply voltages to the stator windings of the synchronous machine, wherein it superimposes fundamental wave components of the voltages applied at low speeds of the synchronous machine with higher-frequency additional signals for operating the synchronous machine , with measuring devices designed and arranged to measure currents flowing through the stator windings in response to the applied voltages, and with an observer for estimating a rotor position of the rotor on the basis of the voltages applied to the stator windings and the currents flowing through the stator windings the observer is designed as an Extended Kalman Filter (EKF), the EKF containing a model of the synchronous machine with a fundamental wave model component and an additional signal model component. The observer or the EKF is responsible for continuously receiving the fundamental wave components and the additional signals applied voltages and the measured currents and arranged for outputting the estimated rotor position to the voltage regulator. The EKF is the only observer who outputs the rotor position to the voltage regulator. The EKF therefore also specifies the rotor position for the voltage regulator at the low speeds, and in the permanently excited synchronous machine according to the invention there is no speed-dependent switching or other synchronization between different observers for estimating the rotor position.

As with the method according to the invention, the superposition of the higher-frequency additional signals at the low speeds of the synchronous machine does not exclude that such a superposition also takes place at higher speeds of the synchronous machine. Furthermore, the fundamental wave components and the additional signals of the applied voltages can be communicated to the EKF in various ways, in particular in the form of specifications for the voltages to be applied, but potentially also differently, e.g. B. in the form of measured values of the applied voltages.

The EKF of the permanent magnet synchronous machine according to the invention is preferably also designed to estimate fundamental wave current components of the currents that flow through the stator windings as responses to the fundamental wave components of the applied voltages and to evaluate them with regard to the rotor position, and is arranged to calculate the estimated fundamental wave current components. to output components together with the estimated rotor position to a regulator of the voltage regulator, which regulates the fundamental current components. The EKF separates the additional current components of the currents that flow through the stator windings as a response to the higher-frequency additional signals of the applied voltages from the measured currents, so that the measured currents are divided into fundamental wave current components and additional current components without any loss of information or phase-shifting filtering. These additional current components are estimated by the EKF and evaluated with regard to the rotor position. In order to estimate all relevant state variables of the permanent magnet synchronous machine, the EKF can furthermore be designed to estimate an angular speed of the rotor, whereby it is then arranged to output the estimated angular speed to the controller, which determines the speed of the synchronous machine by setting suitable basic values - regulates the proportions of the applied voltages, target currents or target torques. An additional signal presetting device of the permanent magnet synchronous machine according to the invention can superimpose an additional signal presetting for the additional signals of a voltage manipulated variable at the output of the controller for the fundamental wave current components. The additional signal specification can have the form of a rotating voltage vector. The regulator is the part of the voltage regulator which applies the voltages to the stator windings and which regulates the fundamental current components.

The EKF can furthermore be designed to estimate a portion of the flux linkage of permanent magnets located on the rotor that induces voltage in the stator windings and / or a mechanical disturbance variable. The EKF can be arranged to transmit the results of these estimates to the voltage regulator. In this way, the voltage regulator can also take these state variables and system parameters into account when regulating the fundamental current components. However, this consideration often turns out to be unnecessary in the case of the permanent magnet synchronous machine according to the invention.

In the method according to the invention and the permanent magnet synchronous machine according to the invention, the equations of state of the EKF can describe the fundamental wave model component of the model of the synchronous machine in the following form:

State equations for the estimation of disturbance variables and parameters can have the following form:

The following definition of a fundamental wave inductance matrix L _αβ (φ _el ) applies:

with:

and the following definition of a matrix of the differential inductances L _{αβ, HF} (φ _el ): L _{αβ, HF} (φ _el )

with:

and the following definition of a resistance _{matrix R αβ, HF of} the additional signal model component for the higher-frequency periodic additional signals:

when defining a permanent magnetic field Ψ _{PM, αβ} (φ _el ):

where the vectors are defined as follows:

and where

R _{S is} a phase resistance of one of the stator windings, L _{d is} a fundamental wave inductance in the d direction,

L _{q is} a fundamental wave inductance in q-direction,

L _{d HF is} a differential inductance in the d direction,

L _{q HF is} a differential inductance in the q direction, L dq, HF is a differential coupling inductance, ψ _{PM is} _{a permanent magnetic field of the synchronous machine} (1), ψ _{α is} a flux linkage in the α direction, ψ _{β is} a flux linkage in the β direction,

R _{αα is} a resistance of the additional signal model component in the «direction,

R _{ββ is} a resistance of the additional signal model component in the / ^ direction,

R _{αβ is} a coupling resistance of the additional signal _{model component, i α is} a current in the α direction, iβ is a current in the β direction, u _{α is} a fundamental wave component of the voltages in the direction, u _{β is} a fundamental wave component of the voltages in the / ^ direction is, u _{α, HF is} a higher-frequency periodic additional signal of the voltages in the direction, u _{β, HF is} a higher-frequency periodic additional signal of the voltages in the / ^ direction and φ _{el is} an electrical rotor position.

The equations of state of the EKF can describe the mechanical model part of the model of the synchronous machine in the following form:

whereby

M _{i is} an internal torque, p is a number of pole pairs of the synchronous machine (1),

M _{Stör is} an estimated disturbance torque, φ _{m is} the mechanical rotor position (22) (of interest), ω _{m is} a mechanical angular velocity and

J _{m is} an inertia. In the EKF of the method according to the invention and the permanently excited synchronous machine according to the invention, a covariance matrix of the system noise for one or more system parameters, for example the permanent magnetic field of the synchronous machine, can be varied depending on a current speed of the synchronous machine. For this purpose, a function of the EKF for weighting an estimate of the system parameter x _{SP as a} function of the current speed of the machine can be defined. Specifically, a main diagonal entry q _{SP of} the covariance matrix of the system noise for the system parameter x _SP can be defined as follows:

where q _{min is} a minimum value of the main diagonal _{entry belonging to x SP}

The covariance matrix of the system noise is q _{max is} a maximum value of the main diagonal entry belonging _{to x SP}

Is the system noise covariance matrix, ω _{N1 is} a first limiting angular velocity, ω _{N2 is} a second limiting angular velocity, and is a mechanical angular velocity. Accordingly, the main diagonal _{entry q SP of} the covariance matrix of the system noise belonging to the system parameter x _{SP is set} to a minimum value q _min up to a first limiting angular speed ω _N1 and a maximum value q above a second limiting angular speed ω _N2 _max . A transition is defined between the two limit angular velocities ω _N1 and ω _N2. This transition does not apply if the first limiting angular speed is equal to the second limiting angular speed. Typically, however, the second limiting angular speed is approximately two to four times as great as the first limiting angular speed.

In the method according to the invention and the permanently excited synchronous machine according to the invention, a current dependency of the fundamental wave _{inductances L can be} approximated by a differentiable function L Appr. This can be done in particular as follows using multidimensional third-order polynomials:

where i _{d is} a current in the d direction, i _{q is} a current in the q direction and a ₀ to a _{11 are} approximation parameters.

In the method according to the invention and the permanently excited synchronous machine according to the invention, the current dependency of the differential inductances can also be determined by continuously differentiable functions L _{HF, Appr.} and l _{HF, Appr.} can be approximated. This can be implemented using third-order polynomials as follows:

Main inductances: L _{HF, Appr.} = b ₀ + b ₁ · i _q + b ₂ · i _q ² coupling inductances: l _{HF, Appr.} = c ₀ + c ₁ · i _q + c ₂ · i _q ² + c ₃ · i _q ³ where i _{q is} a current in the q-direction and b ₀ to b ₂ and c ₀ to c _{3 are} approximation parameters.

Advantageous further developments of the invention emerge from the patent claims, the description and the drawings.

The advantages of features and of combinations of several features mentioned in the description are merely exemplary and can come into effect alternatively or cumulatively without the advantages necessarily having to be achieved by embodiments according to the invention.

With regard to the disclosure content - not the scope of protection - of the original application documents and the patent, the following applies: Further features can be found in the drawings - in particular the illustrated geometries and the relative dimensions of several components to one another and their relative arrangement and operative connection. The combination of features of different embodiments of the invention or of features of different patent claims is also possible in a way deviating from the selected back-references of the patent claims and is hereby suggested. This also applies to features that are shown in separate drawings or mentioned in their description. These features can also be combined with features of different patent claims. Features listed in the claims can also be omitted for further embodiments of the invention, but this does not apply to the independent claims of the granted patent. The number of features mentioned in the claims and the description are to be understood in such a way that precisely this number or a greater number than the specified number is present without the need for an explicit use of the adverb "at least". So if, for example, a model part is mentioned, this is to be understood in such a way that exactly one model part, two model parts or even more model parts are present. The features listed in the patent claims can be supplemented by other features or be the only features that the respective method or the respective synchronous machine has.

The reference signs contained in the claims do not restrict the scope of the subject matter protected by the claims. They only serve the purpose of making the claims easier to understand.

BRIEF DESCRIPTION OF THE FIGURES

In the following, the invention is further explained and described on the basis of an exemplary embodiment shown in the figure.

Fig. 1 shows schematically a permanent magnet synchronous machine according to the invention. FIGURE DESCRIPTION

The sensorless permanently excited synchronous machine 1 according to the invention, shown schematically in FIG. 1, comprises an electromechanical structure 2 with a rotor 3 and a stator 4. The rotor 3, which can be on the outside as an alternative to its inner arrangement indicated in FIG. 1, has in FIG permanent magnets, not shown separately, in order to provide permanent magnetic excitation. The stator 4 is provided with stator windings, also not shown separately. Often it is, as in FIG. 1, through three Connection lines 5 to 7 are indicated, around three stator windings that are electrically isolated from one another. An inverter 8 applies voltages to the stator windings via the connecting lines 5 to 7. In response to these voltages, currents flow through the stator windings and via their connection lines 5 to 7. These currents are measured with measuring devices 9 to 11 in the connection lines 5 to 7. The measured currents 19 to 21 are continuously transmitted to an extended Kalman filter 12, which comprises a model 13 of the permanently excited synchronous machine 1 described by equations of state. In addition, a voltage manipulated variable 14, which is specified by a controller 15, and an additional signal specification 16 are continuously transmitted to the EKF 12. The additional signal default 16 is predetermined by an additional signal default generator 17 and superimposed on the voltage manipulated variable 14 at the output of the controller 15. The voltage manipulated variable 14 specifies fundamental wave components of the voltages that are applied to the stator windings of the stator 4 by the inverters 8 connected on the input side to a DC voltage intermediate circuit 18 via the connecting lines 5 to 7. The inverter 8 and the DC voltage intermediate circuit 18 are typically parts of a frequency converter, which is not shown further here. The controller 15, the inverter 8 and the additional signal pre-generator 17 together form a voltage regulator 28 for the voltages applied to the stator windings of the stator 4 of the synchronous machine 1. Due to the additional signal specification 16, the inverter 8 superimposes higher-frequency additional signals on the fundamental wave components at least at low speeds of the synchronous machine 1. The currents 19 to 21 measured with the measuring devices 9 to 11 are the responses both to the fundamental wave components and to the additional signals of the voltages applied to the stator windings of the stator 4. On the basis of the continuously reported values 14, 17 and 19 to 21, the EKF 12 estimates a rotor position 22 of the rotor 3, fundamental current components 23 of the measured currents 19 to 21, as the relevant state variables of the synchronous machine 1, as responses to the fundamental wave components of the applied voltages flow through the stator windings, an angular velocity 24 of the rotor 3 and additional current components 25 of the measured currents 19 to 21, which flow through the stator windings as responses to the higher-frequency additional signals of the applied voltages. In addition, the EKF 12 can estimate further state variables and system parameters 26, such as a portion of the flux linkage of the permanent magnets located on the rotor 3 that induces voltage in the stator windings and / or a mechanical disturbance variable that indicates external influences on the rotor 3, and also these state variables and Send system parameter 26 to controller 15. In particular based on the rotor position 22 and the fundamental wave current components 23, but also on the other transmitted state variables and The controller 15 specifies the voltage manipulated variable 14 for system parameters 24 and 26 in order to regulate the fundamental wave current components 23 which are optimally suited to achieve a desired angular speed 24 of the rotor 3 and thus a specific speed of the synchronous machine 1. The controller 15 can have separate parts for the d and q components of the fundamental current components 23.

The EKF 12 used for estimating the rotor position 22 is based on the model 13 of the synchronous machine 1, which, in addition to electrical model components, can have a mechanical model component describing the electromechanical structure 2. Furthermore, the EKF 12 can include electrical and / or mechanical model components to take into account the specific application of the synchronous machine 1, for example as a drive for a conveyor belt. The EKF 12 is a non-linear observer that can, in principle, estimate the rotor position 22 on the basis of a model without the influence of anisotropy of the inductances of the synchronous machine. However, the information content for the estimation of the rotor position 22 is dependent on the angular speed 24 of the rotor. Therefore, the basic wave components of the voltages applied to the stator windings specified by the voltage manipulated variable 14 are superimposed on the additional signals specified by the additional signal specification 16, which, if anisotropy is present, provide additional information about the rotor position 22 in the currents 19 to 19 flowing and measured as a response to the applied voltages 21 evoke. This information about the rotor position 22 is also evaluated by the EKF 12. This results in an observer for the rotor position 22, which has a high level of accuracy over the entire speed range of the synchronous machine 1 and reduces the dependence on a clear anisotropy of the inductances of the synchronous machine 1. In addition, the EKF 12 can be used to separate the fundamental current components 23 from the additional current components 25 of the measured currents 19 to 21. No filtering of the measured currents, which is usually used, is thus required in order to provide the fundamental wave current components 23 of the measured currents 19 to 21 required by the controller 15. In addition, there is no signal deterioration and no filter-related phase delay in the fundamental wave current components 23 estimated by the EKF 12. Since the EKF 12 also estimates the rotor position 22 at the low speeds and is therefore used as the only observer for the rotor position 22 in the synchronous machine 1, no speed-dependent switching or other synchronization between different observers is necessary, which has a negative influence on such a switching the function of the controller 15 is omitted. The voltage manipulated variable 14 and the additional signal input 16, which are fed to the EKF 12, form a total of four input variables: the fundamental wave voltages u _{α, f} and u _{ß, f} in the alpha and beta directions as well as the alpha and beta components u _{α, c} and u _{β, c of} the additional signals. The additional signals are characterized by their amplitude U _c and frequency ω _c , which are directly impressed, with the frequency

is constant, while the amplitude U _c can be specified as a function of the speed, precisely known:

The total voltage specification 27, which is transferred to the inverter 8, is an additive superposition of both components:

Calculated output variables of the EKF 12, which are compared with measured variables, are total currents, which are made up of fundamental wave current components and higher-frequency additional current components:

The fundamental wave current components and the higher-frequency additional current components result from the calculation of several model components (see below) and a correction from an observer feedback multiplied by a difference between the measured currents 19 to 21 and the calculated output values of the EKF 12 in, i.e. the total currents i _{α , β} . The state vector of the EKF 12 contains the additional current components from an additional signal model component of the model 13 for higher-frequency excitation by the additional signals as well as the fundamental current components from a fundamental wave model component of the model 13 for the excitation by the fundamental wave components of the voltages applied to the stator windings. In addition, the state vector contains the electrical angular velocity and the rotor position, ie the electrical or mechanical rotor position angle, which can be converted into one another using the number of pool pairs of the synchronous machine. Furthermore, other system parameters and disturbance variables, such as B. the permanent magnetic field as a system parameter and a disturbance torque as a mechanical disturbance variable can be estimated. Then the following applies for the state vector:

The model 13 of the EKF can essentially have three model components: the electrical fundamental wave model component for the excitation of the synchronous machine through the fundamental wave components of the applied voltages, the electrical additional signal model component for the higher-frequency periodic excitation of the synchronous machine through the additional signals and the fundamentally optional mechanical model component to take into account the electromechanical Structure 2 of the synchronous machine including torque calculation. _{The output variables, i.e.} the estimated total currents i αβ, here in the alpha / beta coordinate system, are determined from the state variables of the model components. Electrical model of the synchronous machine when excited by the fundamental wave components (fundamental wave model component)

The voltage manipulated variable 14 at the output of the controller 15 is in a typical frequency range from 0 to 400 Hz. The voltage manipulated variable 14 in dq coordinates can be transformed into the alpha / beta coordinate system _{with the help of the estimated electrical rotor position angle φ el and is used as an input variable for} the fundamental wave model component of the model 13 of the synchronous machine 1 is used, as is known in principle to the person skilled in the art. The output quantities of the fundamental wave model component are the estimated fundamental wave current components. The angle information, which is evaluated in this case, can be found in the term of the electromotive force (EMF). Since this term is linearly dependent on the angular speed, this information can only be used from a certain angular speed, which is typically 0.5 to 1% of the rated speed of the synchronous machine 1.

Electrical model of the synchronous machine with higher-frequency periodic excitation by the additional signals (additional signal model component)

As a result of the additional signal specification 16, the additional signals in the form of a higher-frequency alternating voltage with a typical frequency of the order of magnitude of 1,000 Hz or greater are superimposed on the fundamental wave components of the applied voltages. The additional signal model component includes special parameters for the higher-frequency additional signals such as the differential main and mutual inductances as well as a fully occupied resistor matrix. It is generally known that the resulting higher-frequency additional current components can be displayed as current vectors in the alpha / beta coordinate system as follows:

The phase of the higher-frequency periodic additional current components, which is dependent on the rotor position, is due to an angular dependence of the inductances and requires a difference in the inductances in the d and q directions, i.e. anisotropy.

Mechanical model part of the model of the synchronous machine including torque calculation

The input of the mechanical model component is an equation for calculating the internal torque of the synchronous machine from the currents of the electrical model components. The higher-frequency additional current components often have a negligible influence on the torque. In addition, a torque disturbance variable can be estimated, which compensates for mechanical parameter errors, but primarily takes into account friction losses as well as load torques and disturbance torques. The angle estimation can be improved by means of this disturbance variable estimation. The angular acceleration is derived from the torque and, as a result, the angular velocity and the rotor position angle are determined by integration.

For comparison with the measured currents 19 to 21, the current components of the fundamental wave model component and the additional signal model component are required. For this purpose, the additional current components from the additional signal model component are additively superimposed with the fundamental current components from the fundamental wave model component.

The state variables are corrected by means of the difference between estimated and measured currents multiplied by an observer feedback matrix. The rotor position angle, which is initially predicted with the aid of the mechanical model component, can be highly error-prone, depending on the quality of the model. Since the rotor position is contained both in the fundamental wave model component and in the additional signal model component and thus influences the predicted currents that flow in response to the applied voltages, optimizing the state variables with regard to a minimal difference between measured and estimated currents enables a Correction of the estimated rotor position can be made. The fundamental wave model component supplies reliable angle information from the EMF, which, however, is linearly dependent on the angular velocity and is therefore unusable at low speed and when the system is stationary. Primarily in this speed range, the additional excitation of the synchronous machine with the higher-frequency periodic additional signals ensures that the additional signal model component enables the rotor position to be corrected using anisotropy-based rotor position information. Since the rotor position is corrected using both model components, it is possible to estimate the rotor position in the entire speed range without switching or fading between two different estimation methods. In particular, an additional estimate of the permanent magnetic field of the synchronous machine can be useful, since above all the accuracy when estimating the angular velocity is sensitive to this system parameter. If there is a parameter error, there is an offset in the estimation of the angular velocity. When the synchronous machine is at a standstill, the system parameter of the permanent magnetic field is insufficient or not stimulated at all. The corresponding parameter estimate can therefore drift away. The estimate of the angular velocity can be influenced by this effect, and the velocity estimate can be offset. To solve this problem, the weighting of the estimation of the permanent magnetic field within the covariance matrix of the system noise can be defined as a function of the speed. When the machine is at a standstill, the corresponding weighting is zero, so that the system parameter is kept constant. With increasing speed, the covariance of the permanent magnetic field parameter is also increased up to a final value, so that this permanent magnetic field estimate is corrected again by the EKF.

REFERENCE LIST

1 synchronous machine

2 electromechanical structure

3 rotor

4 stator

5 connection cable

6 connecting cable

7 connection cable

8 inverters

9 measuring device

10 measuring device

11 measuring device

12 Extended Kalman Filters (EKF)

13 Model of the synchronous machine

14 voltage manipulated variable

15 controls

16 Additional signal specification

17 Additional signal defaults

18 DC link

19 measured current

20 measured current

21 measured current

22 rotor position

23 fundamental wave current components

24 angular velocity

25 additional power components

26 further state variables and system parameters

27 Total voltage specification

28 voltage regulator

Claims

PATENT CLAIMS

1. Method for detecting a rotor position of a rotor (3) of a permanently excited synchronous machine (1) with

Applying voltages to the stator windings of the synchronous machine (1), with higher-frequency additional signals superimposed on the fundamental wave components of the voltages applied to operate the synchronous machine (1) at low speeds of the synchronous machine (1),

Measuring currents (19-21) flowing through the stator windings in response to the applied voltages,

Defining an observer for estimating the rotor position (22) of the permanently excited synchronous machine on the basis of the voltages applied to the stator windings and the currents flowing through the stator windings as an Extended Kalman Filter (EKF) (12), which is a model (13) of the synchronous machine (1) includes, and continuous input of the applied voltages and the measured currents (19-21) into the EKF (12), characterized in that the model (13) of the synchronous machine (1) has a fundamental wave model component and an additional signal model component that the fundamental wave components and the Additional signals of the applied voltages are continuously entered into the EKF (12) and that the EKF (12) is used as the only observer for the rotor position (22) at all speeds of the synchronous machine (1).

2. The method according to claim 1, characterized in that the model (13) of the synchronous machine (1) has a mechanical model component.

3. The method according to any one of the preceding claims, characterized in that the EKF (12) is used to estimate fundamental wave current components (23) of the currents that flow through the stator windings as responses to the fundamental wave components of the applied voltages and to estimate the rotor position ( 22), and that the estimated fundamental wave current components (23) are used together with the estimated rotor position (22) in order to regulate the fundamental wave current components (23).

4. The method according to any one of the preceding claims, characterized in that the EKF (12) is used to estimate additional current components (25) of the currents that flow through the stator winding as responses to the higher-frequency additional signals of the applied voltages and with regard to the rotor position (22) to be evaluated.

5. The method according to any one of the preceding claims, characterized in that the EKF (12) is further used to estimate an angular velocity (24) of the rotor (3) without differentiating the rotor position (22).

6. The method according to any one of claims 1 to 4, characterized in that an additional signal specification (16) for the additional signals in the form of a rotating voltage vector is superimposed on a voltage manipulated variable (14) for the fundamental wave components of the applied voltages.

7. The method according to any one of the preceding claims, characterized in that the EKF (12) is used to estimate a voltage-inducing portion of the flux linkage of permanent magnets on the rotor (3) in the stator windings and / or a mechanical disturbance variable.

8. Permanent magnet synchronous machine (1) with a rotor (3),

Stator windings, a voltage regulator (28) which is designed and arranged to apply voltages to the stator windings of the synchronous machine (1), the voltage regulator (28) for operating the synchronous machine (1) applied fundamental wave components of the voltages at low speeds of the synchronous machine (1 ) superimposed with higher-frequency additional signals, measuring devices (9-11) which are designed and arranged to measure currents (19-21) flowing through the stator windings as responses to the voltages applied, and an observer who is used to estimate a rotor position (22) of the rotor (3) is designed as an extended Kalman filter (EKF) (12) based on the voltages applied to the stator windings and the currents flowing through the stator windings, the EKF (12) containing a model (13) of the synchronous machine (1) , and the one to continuously receive the created Voltages and the measured currents (19-21) and for outputting the estimated rotor position (22) to the voltage regulator (28), characterized in that the model (13) of the synchronous machine (1) has a fundamental wave model component and an additional signal model component that the EKF (12) is arranged to continuously receive the fundamental wave components and the additional signals of the applied voltages and that the EKF (12) is the only observer who outputs the rotor position (22) to the voltage regulator (28).

9. Permanent magnet synchronous machine (1) according to claim 8, characterized in that the model (13) of the synchronous machine (1) has a mechanical model component.

10. Permanently excited synchronous machine (1) according to claim 8 or 9, characterized in that the EKF (12) is designed to estimate and with regard to fundamental wave current components of the currents that flow through the stator windings as responses to the fundamental wave components of the applied voltages the rotor position (22) and is arranged to output the estimated fundamental current components (23) together with the estimated rotor position (22) to a controller (15) of the voltage regulator (28) which regulates the fundamental current components (23).

11. Permanent magnet synchronous machine (1) according to one of claims 8 to 10, characterized in that the EKF (12) is designed to generate additional current components (25) of the currents which flow through the stator winding as responses to the higher-frequency additional signals of the applied voltages, to estimate and evaluate with regard to the rotor position (22).

12. Permanent magnet synchronous machine (1) according to one of claims 8 to 11, characterized in that the EKF (12) is designed to further estimate an angular velocity (24) of the rotor (3) without differentiating the rotor position (22), and is arranged to output the estimated angular velocity (24) to the or a controller (15) of the voltage regulator (28), which regulates a speed of the synchronous machine (1).

13. Permanently excited synchronous machine (1) according to one of claims 8 to 12, characterized in that an additional signal predictor (17) an additional signal default (16) for the Additional signals in the form of a rotating voltage vector are superimposed on a voltage manipulated variable (14) at the output of a controller (15) of the voltage controller (28), which regulates the fundamental wave current components (23).

14. Permanent magnet synchronous machine (1) according to one of claims 8 to 13, characterized in that the EKF (12) is designed to induce a voltage in the stator windings portion of the flux linkage of permanent magnets on the rotor (3) and / or a mechanical Estimate disturbance variable, and is arranged to output the inducing portion of the flux linkage and / or the mechanical disturbance variable to one or the controller (15) of the voltage regulator (28).

15. Permanent magnet synchronous machine (1), in particular according to one of claims 8 to 14, with

Stator windings and

Devices which are designed and arranged for carrying out the method according to one of Claims 1 to 8.

16. The method according to one of claims 1 to 7 or permanent magnet synchronous machine (1) according to one of claims 8 to 15, characterized in that the equations of state of the EKF (12) the fundamental wave model component in the form:

with Ψ _αβ = L _αβ (φ _el ) i _αβ + Ψ _{PM, αβ} (φ _el ) and the additional signal component in the form:

describe, whereby equations of state for the estimation of disturbance variables and parameters have the form:

when defining a fundamental wave inductance matrix L _αβ (φ _el ):

with:

when defining a matrix of the differential inductances L _{αβ, HF} (φ _el ):

when defining a resistance _{matrix R αβ, HF of} the additional signal model component:

when defining a permanent magnetic field Ψ _{PM, αβ} (φ _el ):

the vectors are defined as follows:

and where

L _{q is} a fundamental wave inductance in q-direction, L _{d, HF is} a differential inductance in the d direction,

L _{q, HF is} a differential inductance in the q direction,

L _{dq, HF is} a differential coupling inductance,

Ψ _{PM is} a permanent magnetic field of the synchronous machine (1),

Ψ _{α is} a flux linkage in the α direction,

Ψ _{β is} a flux linkage in the β direction,

R _{αβ is} a coupling resistance of the additional signal model component, i _{α is} a current in the α direction, i _{β is} a current in the β direction, u _{α is} a fundamental wave component of the voltages in the α direction, uβ is a fundamental wave component of the voltages in the β direction , u _{α, HF is} a higher-frequency periodic additional signal of the voltages in the α-direction k u _{β, HF is} a higher-frequency periodic additional signal of the voltages in the β-direction and φ _{el is} an electrical rotor position.

17. The method or permanent magnet synchronous machine (1) according to claim 16, characterized in that the equations of state of the EKF (12) describe the or a mechanical model part of the model (13) of the synchronous machine (1) in the form:

whereby

J _{m is} an inertia.

18. The method or permanent magnet synchronous machine (1) according to claim 17, characterized in that a covariance matrix of the system noise for a system parameter of the synchronous machine (1) is varied depending on a current speed of the synchronous machine (1) in the EKF (12).

19. The method or permanent magnet synchronous machine (1) according to claim 18, characterized in that a function of the EKF (12) for weighting an estimate of the system parameters is defined as a function of the current speed of the synchronous machine (1).

20. The method or permanent magnet synchronous machine (1) according to claim 18 or 19, characterized in that a main diagonal entry q _{SP of} the covariance matrix of the

System noise for the system parameter x _SP is defined as follows:

Is the covariance matrix of the system noise, ωΝ1 is a first limiting angular velocity, ωΝ2 is a second limiting angular velocity, and ω _{m is} a mechanical angular velocity.

21. The method or permanent magnet synchronous machine (1) according to one of claims 16 to 20, characterized in that a current dependency of the fundamental wave _inductances L is approximated by a continuously differentiable function L Appr.

22. The method or permanent magnet synchronous machine (1) according to claim 21, characterized in that the current dependency of the fundamental wave inductance L is approximated as follows by multidimensional third-order polynomials:

L _Appr. = A ₀ + a ₁ i _d + a ₂ i _d ² + a ₃ i _d ³ + a ₄ i _q + a ₅ i _q ² + a ₆ i _q ³ + a ₇ i _q i _d + a ₈ i _q ² i _d + a ₉ i _q ² i _d ² + a ₁₀ i _q ³ i _d + a ₁₁ ^' i _q ² i _d ³ where i _{d is} a current in the d direction, i _{q is} a current in the q direction and a ₀ to a _{11 are} approximation parameters.

23. The method or permanent magnet synchronous machine (1) according to one of claims 16 to 22, characterized in that the current dependency of the differential inductances by continuously differentiable functions L _{HF, Appr.} and l _{HF, Appr.} is approximated.

24. The method or permanent magnet synchronous machine (1) according to claim 23, characterized in that the current dependency of the HF inductances is approximated as follows by third-order polynomials: