Method for estimating a rotation angle of a rotor of an electric machine, control unit and vehicle

By calculating a transverse voltage deviation and adjusting the estimated flux in the electric machine model, the method addresses inaccuracies in rotor angle estimation due to aging and manufacturing tolerances, achieving precise control with reduced computational effort.

DE102023128086B4Active Publication Date: 2025-07-10SCHAEFFLER TECHNOLOGIES AG & CO KG
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
DE102023128086
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2023-10-13
Publication Date
2025-07-10
Estimated Expiration
2043-10-13

AI Technical Summary

Technical Problem

Existing methods for estimating the rotation angle of a rotor in electric machines are inaccurate due to non-modelable or not sufficiently accurately modelable influences such as aging and manufacturing tolerances, leading to imprecision and high computational effort.

Method used

A method that calculates a transverse voltage deviation and adjusts an estimated flux in the electric machine model using a factor based on the deviation and rotor rotational speed, allowing for more accurate estimation of the rotation angle by adapting to changing parameters.

Benefits of technology

The method provides improved accuracy in estimating the rotation angle by accounting for non-modelable influences, reducing computational effort and ensuring precise control of the electric machine.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method for estimating a rotation angle (φ est ) of a rotor of an electric machine, in particular a permanently excited synchronous machine, in which a transverse voltage deviation (ΔU q ), which is a difference between a measured transverse voltage (U q,real ) and one based on the estimated rotation angle (φ est ) calculated transverse stress (U q,expected ) is calculated and the calculated transverse stress deviation (ΔU q ) into a tracking controller to estimate the rotation angle (φ est ) is entered, whereby a value used in the model of the electric machine for calculating the calculated transverse voltage (U q,expected ) used estimated flow (ψ PM,est ) in the electric machine according to the quadrature voltage deviation (ΔU q ), where the estimated flow (ψ PM,est ) using a factor (6) which is calculated using a quotient of the transverse stress deviation (ΔUq ) and an estimated rotor speed (ω) of the rotor, characterized in that a value of the factor (6) is limited to a predetermined range.
Need to check novelty before this filing date? Find Prior Art

Description

The present invention relates to a method for estimating a rotation angle of a rotor of an electric machine. The invention also relates to a control unit which is designed and programmed to execute the method and to a vehicle which has the control unit.DE 10 2023 107 740 A1 discloses a method for determining the rotor position relative to the stator without using a rotor position sensor.Operating or controlling / regulating an electric machine without the use of a rotation angle sensor has the advantage that the rotation angle sensor and the cabling required for it can be dispensed with. On the one hand, a cost saving can be achieved in this way, since fewer components are required. On the other hand, the installation space that is not required can be used in another way. The function of the rotation angle sensor is substituted by software for which computing power and storage capacity are to be provided in a control unit of a vehicle.A plurality of methods for estimating the angle of rotation of a rotor of an electric machine are known from the prior art. The methods are distinguished into two main classes.In the first main class, when the rotor of the electric machine is rotated and the voltage is sufficient against induced stress, a method based on voltage equations for an electric machine, such as a permanently excited synchronous machine, is applied. The result of the voltage equations is compared with measured voltages. Voltage deviations in a d-q coordinate system are then determined from the difference between the measured voltages and the voltages determined by the model. Then, the rotation angle and the rotor rotation speed are determined so that the voltage deviations are eliminated and minimized, respectively.In "Position and Speed Estimation Algorithm for Permanent Magnet Synchronous Machines Avoiding Nonlinear Magnetic Effects," by M. Brodatzki, J. Sicher, J. Kobl and M. Braun in 2019, 10th International Conference on Power Electronics and ECCE Asia (ICPE 2019-ECCE Asia), 2019, pp. 1-8, DOI: 10.23919 / ICPE2019-ECCEAsia42246,2019,8796980, an iterative method has been proposed, which, starting from a solution for the rotation angle Gamma_[i-1] and the rotor rotational speed Omea_[i-1] from the last calculation step [i-1], additionally determines four voltage errors in d and q by varying both variables, Gamma_[i-1]+delta_Gamma, Gamma_[i-1]-delta_Gamma, Omega_[i-1]+delta_Omega, Omega_[i-1]-delta_Omega. In a calculation step, several iterations with a reduction of delta_Gamma and delta_Omega can then be considered. The method was later modified and simplified to an iteration per calculation step. With the aid of the two-dimensional flossing method, the new solution Gamma_i and Omega_i is then approximately determined. The solution of a nonlinear equation system with two unknowns thus yields the two unknown variables rotation angle and rotor speed.By computing fewer voltage errors and substituting the two-dimensional flossing method with two linear interpolation problems, a computational effort can be further reduced. The calculation of the rotor rotational speed is based on the equation delta_uq=uq_meas-R*lq-Lq*diq / dtomega*(id*Ld+Psi_em in this case).In order to reduce the voltage errors, a tracking regulator as described in European Patent Specification EP 2 019 482 B1 can also be used. This approach requires the least amount of computing effort of the approaches presented here.The second class of major class are methods using an injection signal. These methods are interesting because they often manage without engine parameters, but depending on the engine they may lead to significant noise problems and should therefore not be used.The methods considered above are dependent on measured currents, on estimated voltages and on parameters of the electric machine.The tracking controller of EP 2 019 482 B1 requires accurate parameters of the electric machine in order to estimate the angle of rotation and the rotor speed appropriately. If the electric machine ages or is operated over a wide temperature range, as is the case for vehicles, it is necessary to track the parameters of the electric machine over time, temperature and current in order to achieve consistent accuracy in the estimation of the angle of rotation. In addition, the parameters may deviate between different electric machines due to manufacturing tolerances. Consequently, when the parameters determined for one electric machine are applied to another electric machine, an imprecision in the estimation of the rotation angle may occur.Moreover, the effort for an accurate determination of the parameters is high. In addition to dependencies on the temperature or the current, which are relatively well known, aging effects and manufacturing tolerances are difficult or hardly mapped and thus lead to unwanted inaccuracies in the estimation of the angle of rotation. Compensation for a change in temperature could be achieved by rotor temperature measurement or estimation. The former cannot be implemented in vehicles for cost reasons and because of the necessary robustness. The latter is frequently used, but is not always sufficiently accurate in the case of calibration despite a high level of complexity.It is therefore an object of the present invention to enable a more accurate estimation of the angle of rotation of a rotor of an electric machine. In particular, it is intended to take account of non-modelable or not sufficiently accurately modelable influences on an accuracy in estimating a rotation angle which is used for controlling or regulating an electric machine.This object is achieved by the method, the control unit and the vehicle having the features according to the independent claims. Advantageous embodiments are the subject of the dependent claims.In a method according to the invention for estimating a rotation angle of a rotor of an electric machine, a transverse voltage deviation is calculated in a model of the electric machine. The electric machine can be in particular a permanently excited synchronous machine. The lateral tension deviation is a difference between a measured lateral tension and a lateral tension calculated based on the estimated rotation angle. The measured transverse voltage can be derived from measured values, in particular from current measured values, and setpoint values, in particular current setpoint values.The calculated lateral voltage deviation is then input to a tracking controller for estimating the rotation angle.The method also adjusts an estimated flux in the electric machine used in the model of the electric machine to calculate the calculated transverse voltage according to the transverse voltage deviation. In this case, a variable which is dependent on the flow in the electric machine can also be adapted.Consequently, the calculated transverse voltage deviation can be appropriately attributed to the magnetic flux, so that it is adapted to the changing parameters of the electric machine. In this way, a more accurate estimate of the rotation angle can be obtained.According to the invention, the estimated flux is adapted on the basis of a factor which is calculated by means of a quotient of the transverse voltage deviation and an estimated rotor rotational speed of the rotor. The factor may be in percent based on a current flow. Consequently, the flow can be adjusted more easily.According to the invention, a value of the factor is limited to a predetermined range. Consequently, the factor is prevented from being excessively adjusted.The adaptation of the flux for a rotor rotational speed between 0 and 100 1 / min can advantageously be prevented. This is important especially when using the factor, since otherwise a division by a value in the vicinity of 0 or even by 0 would take place.According to one aspect, the factor can be set constant. This is particularly useful when short term changes occurring due to temperature or current are replicated in detail in the electric machine model. Consequently, the estimation of the rotation angle is influenced primarily by long-term constant influences, such as the aging of the electric machine and a manufacturing tolerance. The factor determined once can then be stored and reused when the electric machine is restarted. Consequently, an amount of labor in determining the factor is reduced.Advantageously, the constant factor can be established after a predetermined time period has elapsed. The predetermined period of time can be defined in accordance with an occurrence of aging effects. Alternatively or additionally, a determination of the constant factor can be carried out when a predefined threshold value is exceeded by the transverse voltage deviation. This means that aging of the electric machine has occurred. Consequently, it can be ensured that the factor is always matched to the current state of aging of the electric machine.According to another aspect, the factor can be adjusted in a time-variant manner. The time-variant adaptation can be effected by means of a Kalman filter. This is particularly useful when short-term influences, such as temperature and current, are not taken into account in detail in the model of the electric machine. These influences are then taken into account by the time-variant change of the factor.Consequently, a complicated modelling of the electric machine can be avoided.Advantageously, a direct component of the time-variant factor can be used as the initial factor when the electric machine is restarted. The constant proportion of the time-variant factor occurs here on account of aging effects or production tolerances. The factor corresponding to the DC component can then be taken into account during a restart, so that an appropriate estimation of the rotation angle can already take place.The present invention further provides a control unit which is designed and programmed to carry out the method according to one of the preceding aspects. The control unit is therefore designed to receive measured values which are required for estimating the rotation angle and to output the determined rotation angle and / or, if appropriate, a rotor rotational speed derived therefrom. The method according to the invention can also be implemented in a control unit which controls or regulates operation of the electric machine. Consequently, the same advantages as for the method according to the invention can be achieved for the control unit.The present invention also provides a vehicle having the control unit according to the preceding aspect.The present invention will be described in detail below with reference to the figures. The following are shown: FIG. 1 shows a control circuit known from the prior art, which has a motor model and a tracking controller for estimating a rotation angle of a rotor of an electric machine; FIG. 2 shows time diagrams which show simulated profiles of measured values when using a method according to the invention; and FIG. 3 shows time diagrams which show measured curves of measured values when using the method according to the invention.Next, a first embodiment of the present invention will be described with reference to the figures.FIG. 1 shows a control loop known from EP 2 019 482 B1, which has a motor model, i.e. a model for an electric machine, and a tracking controller for estimating a rotation angle of a rotor of an electric machine. In the model of the electric machine, a longitudinal voltage deviation ΔU d and a lateral voltage deviation ΔU q are calculated and input to the tracking controller as input parameters. The calculated transverse voltage deviation ΔU q is to be used below for estimating a rotation angle of a rotor of the electric machine. The longitudinal voltage deviation ΔU d is independent of the magnetic flux in the electric machine and can therefore be ignored.The lateral voltage deviation ΔU q may be calculated as a difference between an actually measured lateral voltage U q,real and an estimated lateral voltage U q,expected, which is calculated from an estimated rotation angle φ est. The actual transverse voltage U q,real can be derived, as can be seen in FIG. 1, from measured currents in the stator of the electric machine and setpoint current values. Since the actual transverse stress U q,real is based on measured values, it is referred to in the present application for the sake of simplicity as measured transverse stress U q,real. The measured transverse voltage U q,real can be calculated approximately on the basis of the following equation (1). However, since the actual magnetic flux ψ PM is not known, calculation is only theoretically possible.The estimated or expected transverse voltage U q,expected can be calculated on the basis of the following equation (2), since the estimated magnetic flux ψSW,estis used for this purpose:Consequently, the lateral voltage deviation ΔU q can be calculated by the following equation (3):For simplicity, the lateral stress deviation ΔU q can be expressed using the equation (4):Accordingly, the lateral voltage deviation ΔU q approximately corresponds to a product of estimated rotor rotational speed ω and flux deviation Δ ψ PM. In other words, the flux deviation Δ ψ PM corresponds to a quotient of the transverse voltage deviation ΔU q and the estimated rotor rotational speed ω. As a result, an influence of the rotor rotational speed ω or the angular speed can be eliminated. The flux deviation Δ ψ PM calculated thereby is then added to the currently estimated magnetic flux ψ PM,est.The transverse voltage deviation ΔU q calculated in the electric machine model can thus be attributed to the estimated magnetic flux ψ PM,est in the electric machine model. The estimated magnetic flux ψ PM,est can be adjusted by considering the flux ψ PM,est as a state variable in a classic observer. By adapting the magnetic flux ψ PM,est it is possible to take account of influences on an accuracy which are not modelable or not accurately enough in estimating the rotation angle φ est.Alternatively to the direct adaptation of the flux, a scaling factor can also be used for the adaptation of the flux ψ PM,est. The factor may be in percent based on a current value of the magnetic flux ψ PM,est.In the case where time-variant influences, such as temperature or current, are taken into account in detail in the model of the electric machine, it is expedient to set the value of the flux deviation Δ ψ PM or of the factor constant. The determined factor may be stored and loaded from the memory upon a start of the electric machine. After a predetermined period of time has elapsed, which is defined in accordance with an occurrence of aging effects, the flux deviation Δ ψ PM or the factor can then be calculated anew. Alternatively or additionally, the flux deviation Δ ψ PM or the factor can be carried out when a predefined threshold value is exceeded by the transverse voltage deviation ΔU q. In this way, it can be ensured that aging effects are taken into account appropriately.Alternatively to constant feedback, a time-variant feedback can also be used. The time-variant feedback is particularly appropriate if time-variant influences, i.e. temperature and current, are taken into account in less detail or not at all in the model of the electric machine. In this case, a direct component of the time-variant factor can be used as the initial factor upon a restart of the electric machine, since the direct component of the time-variant factor occurs because of aging effects or manufacturing tolerances.In addition, it is advantageous if, at a low rotor speed in the range of 0 1 / min, the adaptation of the flux ψ PM,est rests, since otherwise a division by 0 is carried out in the calculation of the flux deviation Δ ψ PM or of the factor.In addition, the flux deviation Δ ψ PM or the factor should be limited to a predetermined range to prevent excessive matching.FIG. 2 shows time diagrams which show simulated profiles of measured values when using the method according to the invention. The top diagram shows the profile of the rotation angle φ est denoted by the reference symbol 1. In the second diagram from above, the curve of the rotor rotational speed, which is provided with the reference numeral 2, is shown. In the third diagram from above, the course of the factor provided with the reference sign 3 is shown. In the bottom diagram, the curve of the rotational angle error φ est provided with the reference number 4 without the flux adaptation and the curve of the rotational angle error φ est provided with the reference number 5 with flux adaptation are shown. As can be seen in the bottom diagram, by adjusting the estimated magnetic flux ψ PM,est from the transverse voltage deviation ΔU q the estimation of the rotation angle φ est can be improved. At high speed times, after completion of an initialization phase, the rotational angle error can be completely eliminated.FIG. 3 shows time diagrams which actually show measured curves of measured values when using a method according to the invention. The first diagram shows the characteristic of the factor, which is provided with the reference numeral 6. The second graph shows the curve of the rotational angle error provided with the reference numeral 7. The rotational angle error was determined on the basis of a comparison with a measured rotational angle. In the third diagram from above, the curve of the current I d provided with the reference number 8 and the curve of the current I q provided with the reference number 9 are shown. The lowermost diagram shows the profile of the rotor rotational speed, which is provided with the reference numeral 10. In the top diagram, it can be seen that the value of the factor changes relatively strongly over time due to poor calibration of the parameters in the model of the electric motor. In the second diagram from above, it can be seen that an angle error can be reduced to approximately -0.2° at times of high rotational speed.It should be noted that the method according to the invention can be implemented by a control unit. The control unit is accordingly designed and programmed to execute the method according to the invention. The control unit is therefore designed to receive measured values which are required for estimating the rotation angle and to output the determined rotation angle and / or, if appropriate, a rotor rotational speed derived therefrom. The method according to the invention can also be implemented in a control unit which controls or regulates operation of the electric machine. The control unit can be installed in a vehicle.List of reference characters1 Rotation angle 2 Rotational speed 3 Factor 4 Rotation angle error when using the conventional method 5 Rotation angle error when using the method according to the invention 6 Factor 7 Rotation angle error 8 I d9 I q10 Rotational speed

Claims

Method for estimating a rotation angle (φ est) of a rotor of an electric machine, in particular of a permanently excited synchronous machine, in which, in a model of the electric machine, a transverse voltage deviation (ΔU q), which is a difference between a measured transverse voltage (U q,real) and a transverse voltage (U q,expected) calculated on the basis of the estimated rotation angle (φ est) is calculated and the calculated transverse voltage deviation (ΔU q) is input into a tracking controller for estimating the rotation angle (φ est), wherein an estimated flux (ψ PM,est) in the electric machine used in the model of the electric machine for calculating the calculated transverse voltage (U q,expected) is adjusted according to the transverse voltage deviation (ΔU q) wherein the estimated flux (ψ PM,est) is adjusted by means of a factor (6) calculated by means of a quotient of the transverse voltage deviation (ΔU q) and an estimated rotor rotational speed (ω) of the rotor, characterized in that a value of the factor (6) is limited to a predefined range.Method according to claim 1, wherein the adjustment of the estimated flux (ψ PM,est) is suppressed for a rotor speed (ω) between 0 and 100 1 / min.Method according to one of Claims 1 to 2, wherein the factor (6) is fixed constant.Method according to Claim 3, wherein a determination of the constant factor (6) is carried out after a predetermined period of time has elapsed and / or when a predetermined threshold value is exceeded by the transverse voltage deviation (ΔU q).Method according to one of Claims 1 to 2, wherein the factor (6) is adjusted time-variably, in particular by means of a Kalman filter.Method according to Claim 5, wherein a constant proportion of the time-variant factor (6) is used as the initial factor when the electric machine is restarted.A control unit configured and programmed to execute the method of any one of claims 1 to 6.A vehicle comprising the control unit according to claim 7.

Citation Information

Patent Citations

  • Method for determining the rotational parameters of a rotation angle, as well as an electric motor

    DE102023107740A1

  • System for determining the position and speed for a permanent magnet rotor of an electric machine

    EP2019482B1