WO2020016002A1

WO2020016002A1 - Method for estimating the speed and position of a rotor of a wound-rotor synchronous machine

Info

Publication number: WO2020016002A1
Application number: PCT/EP2019/067754
Authority: WO
Inventors: Mohamad KOTEICH; Amir MESSALI; Malek Ghanes
Original assignee: Renault S.A.S
Priority date: 2018-07-16
Filing date: 2019-07-02
Publication date: 2020-01-23
Also published as: FR3083863A1; EP3824541A1; CN112425062A; FR3083863B1

Abstract

The invention relates to a method for estimating the speed and position of a rotor (50) of a wound-rotor synchronous machine (50) powered by a three-phase electric power network, comprising: - a step of injecting, into the three-phase electrical network, a high-frequency voltage signal; - a step (101) of demodulating the currents transformed by the second transformation step (101), comprising high-pass or band-pass filtering, and for determining an estimation error signal (ϵ); - a step of estimating (102) the phase shift (ϕ_comp) produced by the rotor acceleration and by the high-pass or band-pass filtering of the demodulation step (101) to refine the estimation error signal (ε) determined during the demodulation step (101); a step (103) of separating the high-frequency component from the low frequency of the measured currents; the method further comprising a second part (12) for gradually estimating the position, speed and acceleration of the rotor, with gain parameters decoupled from each other, according to the sign of the obtained estimation error.

Description

Method for estimating the speed and position of a rotor of a synchronous machine with a wound rotor

The invention relates to the field of synchronous electric machines with wound rotor.

More particularly, the invention relates to a method for determining the position and the speed of the rotor of a synchronous electric machine with a wound rotor.

To control a synchronous electric machine with a wound rotor (abbreviated as MSRB), it is generally necessary to know the position and the speed of the rotor.

A well-known solution of the prior art consists in installing one or more mechanical position and speed sensors on the mechanical shaft of the machine.

However, these mechanical sensors are expensive, bulky, sensitive to the environment (temperature, noise, mechanical oscillations, electromagnetic compatibility, etc.) and reduce the reliability of the system.

Also, to avoid using mechanical sensors, control methods without a mechanical sensor have been developed to ensure the same or better quality control than that of control with a mechanical sensor.

Usually, these sensorless control methods use methods of estimating the mechanical position / speed, also called software sensors, in a closed loop, based solely on the measurement of the currents.

There are also known methods for estimating the position / speed of the rotor by injecting high frequency signals, as described in document US2004070360 A1, which has the effect of allowing detection that is less dependent on the parameters of the machine.

However, these methods all the same remain dependent on the parameters of the electrical machines, and more particularly for the MSRBs, of the stator inductances seen by the rotor. In addition, these techniques are based on knowledge of the characteristics of the signal injected, such as amplitude and frequency. Also, there is the need for a method of estimating the position / speed more reliable and less dependent on the parameters of the synchronous electric machine with wound rotor.

To this end, a method for estimating the speed and the position of a rotor of a synchronous electric machine with a wound rotor powered by a three-phase inverter is proposed, comprising:

- a step of measuring three-phase currents at the input of the synchronous machine with a wound rotor;

- a step of transforming the three-phase currents measured in a two-phase frame;

- a first part including:

- a step of injecting a high-frequency voltage signal at the input of the machine;

characterized in that the first part further comprises determining a rotor position error value comprising:

- a second step of transforming the measured currents transformed into a two-phase frame by a rotation of TT / 4 radians;

a step of demodulating the currents transformed by the second transformation step comprising a high-pass or band-pass filtering, and making it possible to determine an estimation error signal;

a step for estimating the phase shift produced by the rotor acceleration and by the high-pass or band-pass filtering of the demodulation step to refine the estimation error signal determined during the demodulation step;

a step of separating the high frequency component from the low frequency component of the measured currents; said separation step being independent of a low-pass filtering and allowing the determination of the sign of the error of estimation of the rotor position; the method further comprising a second step-by-step estimation of the position, the speed and the rotor acceleration, with gain parameters decoupled from each other, as a function of the sign of the error of estimate obtained.

Thus, a relatively simple and robust estimate of the position, speed and acceleration of the wound rotor can be obtained, depending on the single sign of the estimation error obtained, which is defined as a function of an error signal calculated from the injection of a high frequency voltage. This makes it possible in particular to obtain values for estimating the position, the speed and the acceleration of the rotor independent of each other, and more particularly calibrated by gains independent of each other.

Advantageously and in a nonlimiting manner, the demodulation step comprises a high-pass filtering of said currents. Thus, demodulation is relatively simple and robust and does not generate any delay in the estimate obtained relative to the rotor position.

Advantageously and without limitation, the step of estimating the phase shift comprises low-frequency filtering. Thus the estimation of the phase shift is relatively simple and efficient.

Advantageously and in a nonlimiting manner, the step of estimating the phase shift comprises a phase locked loop. Thus the estimate of the phase shift is controlled relatively robustly.

Advantageously and in a nonlimiting manner, the step of separating the high frequency component from the low frequency component of the measured currents comprises the calculation of an error signal for estimating the rotor position defined by the equation:

in which l _cn is the amplitude of the negative component of the stator current, w _e the pulsation of the injected high-frequency signal, 4> _comp the estimated phase shift, and Q - Q the rotor position error.

Thus, we can simply determine the sign of the rotor position error from the estimation error signal, which then makes it possible to implement the second part of the method to obtain a simple and robust estimate of the speed, the position and acceleration of the rotor.

Advantageously and without limitation, the second part includes the implementation of at least one low-pass filter. The low-pass filter makes it possible to limit the phenomena of chattering of the sign function of the rotor position error. In particular, said low-pass filter is of order 4. Thus, such a filter does not generate any undesirable effect, such as a phase shift, for the estimation of the speed of the position and of the acceleration of the rotor.

The invention also relates to a device for estimating the speed and the position of a rotor comprising means for implementing a method as described above.

The invention also relates to an electrical assembly comprising a synchronous electrical machine with a wound rotor, and an estimation device as described above.

The invention also relates to a motor vehicle comprising an electrical assembly as described above.

Other particularities and advantages of the invention will emerge on reading the description given below of a particular embodiment of the invention, given by way of indication but not limitation, with reference to the appended drawings in which:

- Figure 1 is a schematic view of a control assembly of an electric machine according to an embodiment of the invention;

- Figure 2 is a schematic view of an estimation method according to an embodiment of the invention;

- Figure 3 is a representation of a step of estimating the phase shift of the stator currents of the method according to the embodiment of Figure 2;

- Figure 4 is a representation of a high-frequency / low-frequency separation step, independent of a low-pass filter of the method according to the embodiment of Figure 2;

- Figure 5 is a view of the second estimation part of the method according to the embodiment of Figure 2; and

FIG. 6 is a representation of the geometric transformations of the currents with respect to a rotor reference frame,

- and Figure 7 is a representation of the tracking algorithm according to equation (12).

Referring to Figure 1, a control assembly of an electric machine 1, for example here an electric motor vehicle 1, comprises a torque setpoint device 2, for example an accelerator pedal 2, for requiring torque from the electric machine.

The torque setpoint resulting from the torque setpoint device 2 is then processed by a current regulator 3, then by an inverter 4, in order to provide a control current adapted to the electric machine 5, here a synchronous electric machine with wound rotor. 5.

In order to allow efficient control of the machine, it is necessary to know the position of its rotor, in other words its angular position relative to the stator, its speed, and advantageously its acceleration. To this end, an estimation process 6 is implemented.

Figures 2 to 6 relate to the same embodiment of the estimation method according to the invention, they will be commented on simultaneously.

The method 6 for estimating the speed and the position of the rotor 50 of a synchronous machine with a wound rotor comprises a step 10 of measuring the three-phase currents, and two parts of the method: a first part 100 of signal processing and demodulation and a second part 200 for estimating the position and the speed as a function of the results of the first part.

First of all, the method implements a step 10 of measuring the three-phase currents i _a , i _b , i _c at the input of the synchronous machine with wound rotor. This step is however not necessarily carried out before the first part 100 of the process, it can also be carried out during the first part 100 of the process, for example before it is necessary to call on the values of three-phase currents measured i _a , i _b , i _c .

Then we proceed to a transformation of the three-phase currents measured i _a , i _b , i _c in a two-phase frame ab.

To this end, a transformation is applied in the rotor reference frame 50, as shown in FIG. 6. Also, from the three-phase currents measured i _a , i _a Œ)

i _b , i _c , we deduce the two-phase current system. measured by lB V ^K )

application of the following equation:

This equation (1) describes the measurement of three-phase currents i _a , i _b , i _c according to a three-phase-two-phase static transformation 13 in a frame ab, here a Concordia transform.

In order to model the high frequency behavior of synchronous machines, we then apply a model based on the following two equations:

dvjj ^s

- A voltage-flux model: vf

- "- dt = ('1)

- A current-flow model:

In which L ₀ = ^{Ld Lq} and L ₀ = ^Ld ₂ ^Lq are respectively the mean and differential inductances of the machine, L _d and L _q are the inductances of the axes d and q of the rotating two-phase frame dq, which is the Park frame , vf and i * represent the three-phase voltages and currents of the machine, respectively seen on the stator and y | is the stator magnetic flux of the machine),

To estimate the position, speed and acceleration of AC machines, we use a technique called pulsing technique with an offset of

on the measurement of the current in the estimated two-phase frame

(d / q), (The estimated benchmark representing the benchmark of the estimated Park). Also the axis coordinate system dm and qm is offset from the injection axis coordinate system d and q by

.

We inject a high frequency voltage on the axis (d / q) and we measure the current on the axis dm and qm

The angular phase shift is shown in particular in a rotor reference 50 with reference to FIG. 6.

The drawing technique makes it possible to inject a high frequency voltage (HF) into the estimated two-phase frame (d, q):

v _a = -V _c sin (ü) _c t), v _a = 0 (3)

Or :

V _c is the amplitude of the HF voltage injected; and

o _c is the pulsation of the HF voltage injected. Referring to Figure 6, the current in the offset frame of

with respect to the injection mark is obtained as follows:

^With

and i ^s _sl which are respectively the amplitude of the positive component l _cp , the amplitude of the negative component l _cn and the fundamental component of the stator current i ^, Q being the position of the rotor, and Q the estimated position of the rotor .

Then a demodulation step 101 of the resulting signal is implemented after the injection of the high-frequency voltage.

To this end, a high-pass filter, abbreviated as HPF from the English High-Pass Filter, or according to an alternative, a band-pass filter, abbreviated as SFF from the English Single-Frequency Filtering, is used to filter the current. i ™ in the phase shifted frame so as to remove the fundamental component.

The resulting high frequency current (i ™ _HF ) is obtained according to the following equation:

lcn COS ((2 (0— q) + 5

cos (o) _c t) (5)

l _cn sin ((2 (e - §) + 5

Which gives by trigonometric development

Icp g f Icn [cos ((2 (0 - Q) - sin (2 (0 - §))]

cos (o) _c t) (6)

Icp Y - lcn [sin (2 (q - q)) + cos (2 (0 - Q))]

In this development, the difference i ™ _HF - i ™ _HF is used to extract the position estimation error signal (q - Q). The extraction 101 of the position estimation error signal (q - Q) corresponding to the demodulation 101 of the signal.

[sin (2 (0 - q))] cos (ü) _c t) (7)

The estimation error signal e is formalized according to equation (7), but the angular error q - Q between the position of the rotor and the estimated position of the rotor is a function of this estimation error signal e . Also, by the analysis of the estimation error signal e it will be possible, as described below, to deduce the sign of the position error Q - Q as a function of the sign of the estimation error signal e . The sign of the position error q - Q making it possible to determine in the second part of the process the estimate of the position, the speed and the rotor acceleration.

Once the estimate error signal e has been obtained, according to equation (7), in other words, once the demodulation has been carried out, a phase shift estimation step 102 is implemented, as shown in FIG. 3.

The variations in speed during the acceleration phases of the machine generate a phase shift 4> _acc at the level of the signal carrier (cos (o) _c t + 4> _acc )).

The use of a high-pass filter HPF in the demodulation step 101, or according to an alternative a band-pass filter, abbreviated as SFF from the English Single-Frequency Filtering, also produces a phase shift f _HBR at the level of the carrier (cos (o) _c t + f _HBR )).

Consequently, the carrier signal undergoes these delays and its expression, as formulated previously in equation (7) becomes:

6 A Cos (ü ⁾ _c t + 4> comp) (8)

3VeC

Faoihr FI-IBF F 4 acc- To extract the error of estimation of the position located in the term A, we multiply e by the term cos (ü) _c t + $ _comp ), hence the need to estimate the phase shift f _eoihr .

By multiplying the estimation error of equation (8) by the term sin (ü _c t + $ _comP ) we obtain:

by applying a low-pass filter (abbreviated to LPF from the English Low-Pass Filter), we obtain:

And by applying to (11) a tracking algorithm, here a phase locked loop, abbreviated as PLL from the English Phase Loop Lock, we can calculate the estimate of the phase shift 4> _comp ^:

PLL (LPF (e * sin ü) _c t + Faoihr))) ^- PLL (- sin J _comp - Faoihr)) ^- Fcomp

(12)

This estimate of the phase shift f _e0IΏ r then makes it possible to obtain:

Feoihr ^- Fcomp

And so,

[cos ü) _c t + F _aoihr )]

Indeed, the objective of the phase shift estimation is to reconstruct the signal of the high frequency carrier cos (ü) _c t + $ _CO mp) to obtain the square of this component, square of the carrier (high frequency) [cos (ü) _c t +

is an unknown quantity.

The preceding calculations make it possible to estimate the phase shift f _e0IΏr , which is equal to $ _CO mp once the estimator has converged.

Also, the error of estimation of the phase shift is sent to stages of tracking and optimization of error (PLL), with reference to Figure 7 to make converge f ^ _ihr to F _aoihr ·

Next, a step 103 of separating the high frequency component from the low frequency is implemented, making it possible to avoid the use of low-pass filter (LPF). We then speak in the following description of the step for removing LPF 103.

The estimation error e containing the position of the machine was determined previously so that it can be expressed according to the following equation:

e = -V21 _cn [sin

By multiplying the latter with the term containing the phase shift cos (o) _c t + f _eopir ). ^we get:

[cos (ü) _c t + $ comp)] ²

(14) In the case of protruding wound rotor machines, say protruding machines, [cos (ü) _c t + $ _CO mp)] ² > 0 and - I _cn > 0 because L _q > L _d .

Also :

sign (e * cos (ü) _c t + $ _CO mp) = sign (-V2I _cn [sin (2 (q - §))] [cos (a> _c t + $ _CO mp)] ² )

= sign (- l _cn V2) * sign [sin (2 (q - §))] * sign ([cos (a> _c t + $ _CO mp)] ² )

= sign ([sin (2 (q - q))] (15)

with:

sign [sin

where the term sign means "sign of the expression contained".

The expression according to equation (15) describes the sign of the position estimation error without requiring conventional techniques based on low pass filters (LPF).

This estimation error according to equation (15) is then injected as information into a set of tracking steps 200 according to the invention, step by step and with convergence in finite time.

This set of steps 200 corresponds to the second part 200 of the method according to the invention, which aims to estimate the position, the speed and the acceleration of electric AC machines.

In this second part 200, to make the procedure for adjusting the technique of estimating position, speed and acceleration easy, an estimator, also called observer, robust and step-by-step, as shown in FIG. 4 is put implemented to converge the states of position, speed and acceleration one by one, independently of the other states. This makes it possible to adjust the convergence of these states in finite time, each state being taken separately.

When the estimated position Q is considered to be equal to the real position Q according to equation (19), in other words when we consider, by approximation that the error q - Q of position is zero: If q ”q (19)

so

sin (0 - §) = ø - § (20)

The only measure for the estimator is:

o (t) = sign (e - q), from the first part 100 of the process.

The robust step-by-step observer proposed to estimate position, speed and acceleration is defined by the following equations:

q = ώ + K ₀ sign (0— q) (21)

ώ = a + E ₁ K _Cù sign (ô) - ώ) (22)

a = E ₂ K _a sign (â— a) (23)

OR :

w = ώ + Ke [sign (0 - §)] Filtered (24)

â = a + K _w [sign (ü) - S)] _Filtered (25) with

OR

s (z) = TZ (s (ί)), s = sign (0— q) (29)

where TZ represents the Transform in Z which transforms the time function o (t) into the discrete function s (z).

The function f (z) is introduced so as to detect the phenomenon of browsing, since only the sign of the estimation error is available as information for the observer; the rotor position is not available for measurement.

In order to obtain the filtered values of speed w and acceleration a, we apply low-pass filters (LPF) of order 4 used in the second part 200 of the method, with reference to FIG. 5. These low-pass filters are introduced to reduce the chatter phenomena of the sign function and do not influence the position, speed and acceleration estimates because these estimates are advantageously decoupled from each other.

The virtual mechanical system of position, speed and acceleration used for the design of the observer (21), (22) and (23) is given as follows:

q = w (30)

ώ = a (31)

ά = 0 (32)

Equations (33), (34) and (35) define the error in estimating the position, speed and acceleration between equations (30) - (31) - (32) and the observer (21 ) - (22) - (23)

e ₀ = q— Q (33)

q _w = w— ώ (34)

e _a = a— a (35)

We deduce the estimation error dynamics by the following equations:

è ₀ = q _w - Kesign (0 - Q) (36)

έ _w = e _a - E ₁ K _Cù sign (ô) - ώ) (37)

è _a = - E ₂ K _w sign (a— a) (38)

with K _q > Max (| e _w |), K _w > Max (| e _a |) and K _a > 0 gains defining a positive value to limit the noises.

Thus, it appears that the method according to the invention ensures the convergence of the error of estimation error of the position, the speed and the acceleration (36) - (37) - (38) to zero in finite time.

Claims

1. Method for estimating the speed and the position of a rotor (50) of a synchronous electrical machine with a wound rotor (50) supplied by a three-phase inverter, comprising:

- A step of measuring three-phase currents (i _a , i _b , i _c ) at the input of the synchronous machine with a wound rotor (50);

- a step of transforming the three-phase currents (i _a , i _b , i _c ) measured in a two-phase frame (dq),

- a first part (11) comprising:

characterized in that the first part (11) further comprises determining a rotor position error value comprising:

- a second transformation step (101) of the measured currents transformed into a two-phase frame (dq) by a rotation of p / 4 radians;

- a demodulation step (101) of the currents transformed by the second transformation step (101) comprising high-pass or band-pass filtering, and making it possible to determine an estimation error signal (e);

a step of estimating (102) the phase shift (4> _com p) produced by the rotor acceleration and by the high-pass or band-pass filtering of the demodulation step (101) to refine the error signal estimation (e) determined during the demodulation step (101);

a step of separation (103) of the high frequency component from the low frequency component of the measured currents; said separation step (103) being independent of low-pass filtering and allowing the determination of the sign of the error in estimating the rotor position;

the method further comprising a second part (12) of step-by-step estimation of the position, the speed and the rotor acceleration, with gain parameters decoupled from each other, according to the sign of the estimation error obtained.

2. Method according to claim 1, characterized in that the demodulation step (101) comprises a high-pass filtering of said currents.

3. Method according to claim 1 or 2, characterized in that the step of estimating (102) the phase shift comprises low-frequency filtering.

4. Method according to claim 3, characterized in that the step of estimating (102) the phase shift comprises a phase locked loop.

5. Method according to any one of claims 1 to 4, characterized in that the step of separation (103) of the high frequency component of the low frequency of the measured currents comprises the calculation of an error signal of estimation of the rotor position defined by the equation:

e = - V2I _cn [sin

in which l _cn is the amplitude of the negative component of the stator current, w _e the pulsation of the injected high-frequency signal, and 4> _comp the estimated phase shift (102) and (q - Q) the rotor position error.

6. Method according to claim any one of claims 1 to 5, characterized in that the second part (12) comprises the implementation of at least one low-pass filter.

7. Method according to claim 6, characterized in that said low-pass filter is of order 4.

8. Device for estimating the speed and the position of a rotor comprising means for implementing a method according to any one of claims 1 to 7.

9. An electrical assembly comprising a synchronous electrical machine with a wound rotor, and an estimation device according to claim 8.

10. Motor vehicle comprising an electrical assembly according to claim 9.