Method for determining the angular offset between the rotor and the stator of an electrical machine of a motor vehicle

Abstract

The invention relates to a method for determining the angular offset between the rotor and the stator of an electrical machine for driving a motor vehicle, the rotor being supplied with a rotor-energizing current and voltage, the stator being supplied over three phases with stator-phase currents and voltages, wherein the method includes the following steps: verifying that the electrical machine is stopped; applying a rotor-energizing signal for an amount of time capable of causing the partial magnetization of the rotor; applying a rotor-de-energizing signal capable of causing an active and rapid demagnetization of the rotor, while measuring, during the magnetization, the stator-phase currents and maintaining zero voltage among the phases of the stator; determining the direct stator currents and quadrature stator currents corresponding to the stator-phase currents measured by applying Park and Clarke transforms and in accordance with a measurement of the electrical position of the rotor; determining the maximum values among the direct and quadrature stator currents; transmitting an angular offset fault signal if the absolute value of the maximum quadrature stator signal is greater than a threshold value; determining a correction of the angular offset in radians by calculating the tangent arc of the maximum quadrature stator current value divided by the maximum value of the direct stator current.

Claims

1. A method for determining the angular offset between the rotor and the stator of an electrical machine, the rotor being supplied with a rotor excitation current and voltage, the stator being supplied on three phases with stator phase currents and voltages, characterized in that it comprises the following steps: verifying that the electrical machine is at rest, applying a rotor excitation signal for a duration capable of causing partial magnetization of the rotor, applying a rotor deexcitation signal capable of causing active and rapid demagnetization of the rotor, while measuring the stator phase currents during the demagnetization, and while maintaining zero voltages between the phases of the stator, determining direct stator currents and quadrature stator currents, corresponding to the stator phase currents measured by applying the Park-Clarke transformation and as a function of a measurement of the electrical position of the rotor, determining the maximum values among the direct stator currents and the quadrature stator currents, transmitting an angular setting fault signal if the absolute value of the maximum value of the quadrature stator current is greater than a threshold value, determining an angular offset correction in radians by calculating the arctangent of the ratio of the maximum value of the quadrature stator current and the maximum value of the direct stator current.

2. The method as claimed in claim 1, wherein partial magnetization of the rotor is carried out by causing the appearance of a rotor excitation current greater than a magnetization threshold current.

3. The method as claimed in claim 1, wherein the demagnetization of the rotor is carried out by causing a reduction of the rotor excitation current below a demagnetization threshold value.

4. The method as claimed in claim 1, wherein the rotor excitation signal and the rotor deexcitation signal are rotor excitation currents or rotor excitation voltages.

5. The method as claimed in claim 1, wherein the rotor excitation signal and the rotor deexcitation signal form a pattern selected from among patterns of rectangular, triangular, sinusoidal or Dirac shape.

6. The method as claimed in claim 1, wherein the maximum values of the direct stator currents and of the quadrature stator currents at the end of demagnetization are determined, when the rotor excitation current is zero.

7. The method as claimed in claim 1, wherein the maximum values of the direct stator currents and of the quadrature stator currents are determined at the start of magnetization, when the excitation current reaches an adjustable threshold.

Description

(1) Other objects, characteristics and advantages will become apparent on reading the following description, given solely by way of nonlimiting example and provided with reference to the appended figures, in which:

(2) FIG. 1 illustrates the main steps of the method, and

(3) FIG. 2 illustrates the various reference coordinates and angles of an electrical machine.

(4) As seen in the introduction, an electrical machine comprises a stator, the fixed part of the machine, and a rotor, the rotating part connected to the output shaft and making it possible to transmit the mechanical torque produced.

(5) The stator comprises windings, through which the stator currents flow and which produce magnetic fluxes guided by the body of the stator in order to be closed on the rotor.

(6) The rotor comprises windings, through which the rotor excitation current flows and which produce excitation magnetic fluxes which are guided by the body of the rotor in order to be closed on the stator.

(7) The automobile motors to which the present invention relates are synchronous machines having a wound rotor.

(8) According to the prior art, a vector representation of the fluxes and currents according to a Fresnel diagram provides a good model of the machine. The reference frame adopted is a rotating reference frame fixed to the rotor. The axis of the rotor winding is denoted as direct, d. The axis in quadrature is denoted as q.

(9) The stator currents are represented by a current vector {right arrow over (i.sub.s)}, a fixed vector with components i.sub.d and i.sub.q in the rotating reference frame ({right arrow over (d)}, {right arrow over (q)}j). The direct stator current i.sub.d and the quadrature stator current i.sub.q are obtained on the basis of the measurements of the phase currents of the stator (i.sub.A, i.sub.B and i.sub.c) and of the electrical position by means of the Park-Clarke transformation. The rotor excitation current is represented by a vector {right arrow over (i.sub.f)} collinear with the axis d.

(10) The stator {right arrow over (i.sub.s)} and rotor {right arrow over (i.sub.f)} currents produce a magnetic flux {right arrow over (.sub.s)}, through the windings of the stator and {right arrow over (.sub.f)} through the winding of the rotor. The stator currents are generated by a voltage vector {right arrow over (V.sub.s)} resulting from the voltages between phases produced by a control system. The rotor excitation current is generated by the rotor deexcitation voltage V.sub.f produced by the excitation control system.

(11) The vector electrical equations of the machine are as follows:

(12) $\begin{matrix} \overset{.fwdarw.}{V_{s}} = R_{s} .Math. \overset{.fwdarw.}{j_{s}} + \frac{\overset{.fwdarw.}{_{s}}}{t} & (Eq . 1) \end{matrix}$
for the stator; and

(13) $\begin{matrix} V_{f} = R_{f} .Math. i_{f} + \frac{_{f}}{t} & (Eq . 2) \end{matrix}$
for the rotor,
with R.sub.s being the phase-neutral stator resistance and R.sub.f the rotor winding resistance.

(14) Since the current vector {right arrow over (i.sub.s)} can be decomposed into currents i.sub.d and i.sub.q, the voltage vector {right arrow over (V.sub.s)} can also be decomposed similarly into components V.sub.d and V.sub.q, the flux {right arrow over (.sub.s)} also being decomposed into fluxes .sub.d and .sub.q.

(15) The following system of equations is then obtained:

(16) $\begin{matrix} {\begin{matrix} V_{d} = R_{s} .Math. i_{d} + \frac{_{d}}{i_{d}} .Math. \frac{i_{d}}{t} + \frac{_{d}}{i_{f}} .Math. \frac{i_{f}}{t} - .Math._{q} \\ V_{q} = R_{s} .Math. i_{q} + \frac{_{q}}{i_{q}} .Math. \frac{i_{q}}{t} + .Math._{q} \\ V_{f} = R_{f} .Math. i_{f} + \frac{_{f}}{i_{f}} .Math. \frac{i_{f}}{t} + \frac{_{f}}{i_{d}} .Math. \frac{i_{d}}{t} \end{matrix} & (Eq . 3) \end{matrix}$
with the angular frequency of the machine (angular velocity multiplied by the number of pole pairs).

(17) At low currents, since the flux levels are low and the machine is not magnetically saturated, these equations can be linearized. The following equations are then obtained:

(18) $\begin{matrix} {\begin{matrix} V_{d} = R_{s} .Math. i_{d} + L_{d} .Math. \frac{i_{d}}{t} + M_{f} .Math. \frac{i_{f}}{t} - .Math. L_{q} .Math. i_{q} \\ V_{q} = R_{s} .Math. i_{q} + L_{q} .Math. \frac{i_{q}}{t} + .Math. L_{d} .Math. i_{d} \\ V_{f} = R_{f} .Math. i_{f} + L_{f} .Math. \frac{i_{f}}{t} + \frac{3}{2} .Math. M_{f} .Math. \frac{i_{d}}{t} \end{matrix} & (Eq . 4) \end{matrix}$

(19) In particular, at zero speed (=0) and if V.sub.d=V.sub.q=0, Equation 4 becomes

(20) $\begin{matrix} {\begin{matrix} i_{d} = - \frac{1}{R_{s}} .Math. (L_{d} .Math. \frac{i_{d}}{t} + M_{f} .Math. \frac{i_{f}}{t}) \\ i_{q} = - \frac{L_{q}}{R_{s}} \frac{i_{q}}{t} \end{matrix} & (Eq . 5) \end{matrix}$

(21) It is to be noted that i.sub.q does not depend on i.sub.f. A variation in i.sub.f therefore alters i.sub.d but not i.sub.q.

(22) The determination method therefore consists in carrying out a rapid variation in current i.sub.f when the machine is at rest (co=0) and the stator voltage is zero ({right arrow over (V.sub.s)}={right arrow over (0)}). This may be obtained by controlling the voltage V.sub.f according to a pattern, for example a step pattern. The effect of this control is magnetization of the rotor followed by rapid demagnetization thereof, which causes the appearance of an induced current in the stator.

(23) By integrating the differential equations for the currents (Eq. 5), for the currents i.sub.d, i.sub.q obtained on the basis of the measured phase currents, after a Park-Clarke transformation, an invariant current i.sub.q is obtained in the case in which the position measurement offset is correctly learned, and a current i.sub.q proportional to i.sub.d in the converse case. It is then possible, by observing the variation in the current i.sub.q, to carry out a diagnosis of the determination of the offset, for example by comparing the current i.sub.q with a stored detection threshold: if the current i.sub.q exceeds this threshold during the demagnetization, the activation of an angular setting fault signal is triggered.

(24) The real value of the position measurement offset ( in degrees of arc) can then be obtained by taking the peak current values of i.sub.d and i.sub.q, denoted as i.sub.d.sub._.sub.dr and i.sub.q.sub._.sub.dr at the time of the rapid demagnetization of the rotor, denoted as instant t d.sub.r, and applying the following formula:

(25) $\begin{matrix} = \frac{180}{} .Math. \arctan (\frac{i_{q_dr}}{i_{d_dr}}) . & (Eq . 6) \end{matrix}$

(26) The method for determining the angular offset between the rotor and the stator comprises the following steps, which are illustrated by the single figure.

(27) During a first step 1, the method is initialized. To this end, the fact that the absolute value of the rotation speed is less than a threshold .sup.coseu.sup.ii is verified. Alternatively, the method may be initialized each time the control system of the machine is initialized, or only under certain triggering conditions, for example after a maintenance operation.

(28) During the second step 2, the application of an excitation voltage V.sub.f is controlled for a duration capable of causing partial magnetization of the rotor, that is to say causing the appearance of a current i.sub.f greater than a magnetization threshold current. For example, V.sub.f=3 0 V may be applied for 0.3 s.

(29) Immediately after the end of the duration of step 2, during a third step 3 a new voltage V.sub.f capable of causing active and rapid demagnetization of the rotor, that is to say reduction of the current i.sub.f below a demagnetization threshold value, is applied. For example, V.sub.f=V.sub.bat may be applied, with V.sub.bat being the battery voltage.

(30) The rapid variation in i.sub.f induces a current in the stator only along the axis d, as predicted by Equation 5.

(31) The square-wave variation applied to the voltage V.sub.f, or to the current setpoint i.sub.f, may be replaced with any pattern consisting of a variation in the excitation current capable of causing induction of a current in the stator. Triangular, sinusoidal or Dirac shape variations may be mentioned in particular.

(32) Alternatively, the magnetization and the demagnetization of the rotor may be carried out by controlling the excitation current i.sub.f rather than the voltage V.sub.f. This has the advantage of better reproducibility of the current profiles, while accelerating the determination of the offset.

(33) Furthermore, the stator phase currents i.sub.A, i.sub.B and i.sub.C are measured during the demagnetization period.

(34) During steps 2 and 3, voltages are applied between zero phases of the stator in order to obtain looping by the phases of the induced stator currents.

(35) During a fourth step 4, the Park-Clarke transformation is applied to the measured currents, as a function of the electrical position measurement of the rotor .sub.e in order to obtain the measured Park currents i.sub.d and i.sub.q.

(36) It is to be recalled that the expression of the Park-Clarke transformation is:

(37) $\begin{matrix} (\begin{matrix} i_{d} \\ i_{q} \end{matrix}) = \frac{2}{3} .Math. [\begin{matrix} \cos (_{e}) & \cos (_{e} - 2 / 3) & \cos (_{e} - 4 / 3) \\ - \sin (_{e}) & - \sin (9_{e} - 2 / 3) & - \sin (9_{e} - 4 / 3) \end{matrix}] .Math. (\begin{matrix} i_{A} \\ i_{B} \\ i_{C} \end{matrix}) & (Eq . 7) \end{matrix}$

(38) Still during the fourth step, the maximum values among the measurements of the currents i.sub.d and i.sub.q are determined. These maximum values are measured as values i.sub.d.sub._.sub.dr and i.sub.q.sub._.sub.dr. The maximum currents are reached when the gradient of the rotor excitation current is at its maximum. This may be the case at the end of demagnetization, for example, when the current i.sub.f reaches zero.

(39) During a fifth step 5, the value of i.sub.q.sub._.sub.dr is compared with a threshold value i.sub.q.sub._.sub.max.

(40) If the comparison |i.sub.q.sub.dr|<i.sub.q.sub.max is satisfied, there is no setting error. The method ends with a step 6 which consists in approving the control system, allowing the functional mode of the motor to be entered.

(41) In the converse case, an angular setting fault signal is transmitted. This fault expresses a poor angular setting of the measured absolute angular position with respect to the actual position of the rotor with respect to the stator. The consequence of this poor setting would be poor phasing of the stator currents, and therefore of the stator flux, which would induce poor production of the mechanical torque, or even reversal of the running direction of the machine.

(42) The method then continues with step 7, during which a correction of the angular offset is determined by applying Equation 6. The correction value of the offset may be transmitted. Alternatively, during an eighth step 8, the corrected value .sub.0.sub._.sub.corr of the position measurement offset .sub.0 is determined in the following way:

(43) $\begin{matrix} _{0_corr} =_{0} - \frac{180}{} .Math. \arctan (\frac{i_{q_dr}}{i_{d_dr}}) & (Eq . 8) \end{matrix}$

(44) The corrected offset .sub.0.sub.corr is applied in the following way:
.sub.e=mod(0.sub.raw.sub.0.sub._.sub.corr,360)(Eq. 9)
with
.sub.raw being the raw position measurement provided by the measurement device without angular setting, and
.sub.e being the value of the absolute angular position.

(45) It should be noted that the correction of the offset .sub.0 is calculated on the basis of the measurements of stator currents induced during the demagnetization. It is also possible to calculate the correction of the offset on the basis of the stator currents measured during the magnetization. In this case, the maximum values of these currents are obtained at the start of the sequence, when the gradient of the current i.sub.f is maximum, or, for example, at the time when it crosses an adjustable predetermined threshold.

Method for determining the angular offset between the rotor and the stator of an electrical machine of a motor vehicle

Assignee

Inventors

Cpc classification

Classification Explorer

H02P21/18

ELECTRICITY

Classification Explorer

G01D5/12

PHYSICS

Classification Explorer

H02P21/32

ELECTRICITY

Classification Explorer

H02P6/16

ELECTRICITY

Classification Explorer

H02P2203/03

ELECTRICITY

Classification Explorer

G01B7/30

PHYSICS

International classification

Classification Explorer

G01R13/00

PHYSICS

Classification Explorer

H02P21/32

ELECTRICITY

Classification Explorer

G01D5/12

PHYSICS

Classification Explorer

H02P6/16

ELECTRICITY

Classification Explorer

G01R15/00

PHYSICS

Classification Explorer

H02P23/00

ELECTRICITY

Classification Explorer

G01B7/30

PHYSICS

Classification Explorer

H02P21/18

ELECTRICITY

Abstract

Claims

Description