Method and apparatus for estimating the electromagnetic torque of a synchronous motor

Abstract

An apparatus for estimating the electromagnetic torque of a three-phase synchronous motor including permanent magnets, comprising: a first flux estimator (10) including two electromotive force integrators (103) for estimating corresponding components of the flux in a fixed two-phase coordinate system associated with the stator; an estimator (20) for estimating the torque from the corresponding estimated flux components; a second flux estimator (40) using current expressed in a rotating two-phase reference system associated with the rotor, the second flux estimator having an observer for determining variables characterizing magnetic uncertainties of the motor in order to correct the flux estimation of the second estimator; and a detector (30) for generating a signal (RES) for resetting the initial flux conditions of the two integrators based on the flux estimation performed by the second estimator when the difference between the estimated torque and the setpoint torque is higher than a predefined threshold.

Description

Technical Field

[0001] This invention relates to a method for estimating the electromagnetic torque of a three-phase synchronous motor, particularly a permanent magnet three-phase synchronous motor (PMSM). Background Technology

[0002] This invention is advantageously applicable to the fields of hybrid electric vehicles and electric motor vehicles, wherein this type of electric motor is widely used as a traction motor in the electric traction mode of the vehicle, particularly due to its robustness, simplicity and low maintenance cost.

[0003] In electrically driven motor vehicles, especially for safety and / or control reasons, it is necessary to know the true value of the motor torque provided by the electric motor. Generally, the mechanical drive torque of the electric motor is obtained from a torque meter, which measures the torque on the shaft. However, for many reasons such as cost, reliability, and size, it is preferable to remove the torque meter from the drive system and replace it with a torque estimator.

[0004] Torque estimators are designed to use certain available variables (such as voltage, current, and rotor position) to estimate the true value of the torque provided by the motor.

[0005] Conventionally, an electric motor can be represented using two two-phase models, employing the Clarke transform and the Park transform, respectively. The Clarke transform is based on a fixed two-phase reference frame associated with the stator, while the Park transform is based on a rotating two-phase reference frame associated with the rotor. Therefore, the quantities transformed in the stator and rotor-related reference frames are expressed using exponents α and β on one hand, and exponents d and q on the other, corresponding to the two axes in the fixed two-phase reference frame and the two axes in the rotating two-phase reference frame, respectively.

[0006] Because the magnetic flux of an electric motor is closely related to the torque it provides, the torque of the motor can be observed by estimating the flux. Several techniques are currently known for estimating the flux of a permanent magnet three-phase synchronous motor based on two-phase current and voltage. These techniques are divided into two categories: those based on voltage models and those based on current models.

[0007] Regarding the voltage-model-based flux estimation technique, the electromotive force (EMF) is calculated using voltage and current measured in a reference frame associated with the stator. The stator flux is then obtained by calculating the integral of the EMF.

[0008] However, this estimation method implies a precise understanding of the initial magnetic conditions (flux of the permanent magnet and stator inductance), which is not always possible in reality. Therefore, due to this uncertainty, the estimate of the flux magnitude may be biased relative to the actual flux of the motor. This bias in flux estimation using a voltage-flux model subsequently produces an oscillating term in the estimated torque.

[0009] Furthermore, demagnetization and magnetic saturation, which are prone to occur during motor operation, also produce oscillating terms in the torque estimated using the voltage-flux model of the motor.

[0010] Therefore, another source of deviation in the estimated torque lies in the nonlinearity of the inverter that supplies power to the motor and / or in the measurement uncertainty, which causes a deviation in the calculated electromotive force value and thus in the estimated flux value.

[0011] Another problem to be solved is ensuring observability at low or zero speeds.

[0012] When certain filters are added to address the aforementioned issues (unknown magnetic conditions and electromotive force shift), observability problems arise at low speeds.

[0013] The prior art discloses the following documents related to flux estimation based on voltage models.

[0014] Document US 9985564 B2 describes a method for estimating the stator flux of a permanent magnet three-phase synchronous motor based on a voltage model in two-phase reference frames α and β. Instead of using a pure integrator to integrate the electromotive force, the document proposes using an electromotive force low-pass filter and subsequently adding a compensator to correct for the gain and phase changes introduced by the filter.

[0015] Document CN 103346726 A also proposes estimating the flux of a permanent magnet three-phase synchronous motor based on electromotive force. This document proposes using a second-order filter instead of a pure integrator. It also proposes estimating the q-axis inductance in a two-phase reference frame to improve the estimation performed using a voltage model.

[0016] Document US 7098623 B2 describes a method for estimating rotor position by estimating stator flux using a voltage model in a two-phase reference frame d,q. This document proposes adaptive flux observation based on the injection of a high-frequency signal at low speeds.

[0017] Document EP 1513250 describes a method that combines a voltage model-based flux estimation method with the injection of high-frequency signals to estimate position.

[0018] Regarding current-model-based methods, current, preferably in a reference frame associated with the rotor, is used to estimate the flux. These current-model-based methods overcome the aforementioned limitations associated with voltage-model-based methods because they do not involve the use of open-loop integrators. However, problems to be addressed with these methods include variations in inductance due to magnetic saturation, uncertainties in the flux of permanent magnets, and the need for position sensors.

[0019] The following documents are known from the prior art related to flux estimation based on current models.

[0020] Document US 7774148 B2 describes a method for estimating the torque of a permanent magnet three-phase synchronous motor based on flux estimation using a current model. The fluxes along the d-axis and q-axis are represented by two nonlinear current functions.

[0021] Document CN 103872951 A describes a torque estimator for a permanent magnet three-phase synchronous motor, which is based on flux estimation using a sliding mode observer applied to a voltage model combined with a current model.

[0022] Document JP 2004166408 A describes an adaptive flux estimator for a permanent magnet three-phase synchronous motor, which is based on a voltage model with a correction term based on a current model. Summary of the Invention

[0023] In this context, the object of the present invention is to provide a method for estimating the torque of an electric motor, particularly a permanent magnet synchronous motor in an electric or hybrid motor vehicle, which is robust relative to the aforementioned limitations.

[0024] Therefore, the present invention relates to a method for estimating the electromagnetic torque of a permanent magnet three-phase synchronous motor in an electric or hybrid motor vehicle, the method comprising the following steps:

[0025] - Measure the current and voltage on the phases of the stator of the motor in a fixed three-phase reference frame associated with the stator.

[0026] - Determine the current and voltage in the fixed two-phase reference frame associated with the stator based on the current and voltage measurements in the three-phase reference frame.

[0027] - A first estimate of the motor flux is calculated by integrating the electromotive force, which is based on the voltages and currents expressed in a fixed two-phase reference frame associated with the stator.

[0028] - The electromagnetic torque is estimated based on the flux estimated through the first estimate.

[0029] The method is characterized by comprising the following steps:

[0030] - Determine the currents and voltages in the rotating two-phase reference frame associated with the rotor based on the currents and voltages expressed in the fixed two-phase reference frame and the position of the rotor relative to the stator.

[0031] - A second estimate of the motor flux is calculated in parallel using these currents expressed in a rotating two-phase reference frame associated with the rotor. This second flux estimate is corrected by introducing variables characterizing the magnetic uncertainty of the motor, which are determined by the observer based on these stator currents and stator voltages expressed in a rotating two-phase reference frame associated with the rotor.

[0032] - Compare the estimated torque with the setpoint torque, and

[0033] - When the difference between the estimated torque and the setpoint torque exceeds a predefined threshold, the flux is reset to meet the initial conditions of the first estimate based on the flux estimated by the second estimate.

[0034] Advantageously, the method may include the step of: stopping the calculation of the second flux estimate when the difference between the estimated torque and the setpoint torque is less than or equal to the threshold.

[0035] Advantageously, the method may include the step of determining the percentage of time during which the calculation of the second flux estimate is activated relative to the operating time of the motor.

[0036] Advantageously, the method may include the step of compensating for a bias in the calculated electromotive force before using it in the first flux estimate.

[0037] Advantageously, the second estimation of the flux of the electric motor includes the following steps:

[0038] - Determine the flux-current model of the motor by considering its saturation function, so that the changes in the motor's magnetic parameters caused by magnetic saturation and demagnetization are accounted for in the model.

[0039] - The variables characterizing the magnetic uncertainties of the motor, which are inherent to variations in the motor's magnetic parameters, are introduced into the model.

[0040] - Estimate the variables in the model, which is performed by the observer.

[0041] - Calculate the second flux estimate based on the estimated variables.

[0042] Advantageously, the magnetic parameters under consideration are the linear inductance of the stator on each axis of the rotating two-phase reference frame associated with the rotor and the magnetic flux generated by the permanent magnets of the motor rotor.

[0043] Advantageously, these saturation functions represent magnetic saturation and cross-saturation in the motor.

[0044] Advantageously, the method includes the following steps: using the Kalman algorithm to tune the observer.

[0045] The present invention also relates to an apparatus for estimating the electromagnetic torque of a permanent magnet three-phase synchronous motor in an electric or hybrid motor vehicle, the apparatus being designed to receive currents and voltages measured on phases of the stator of the motor in a fixed three-phase reference frame associated with the stator, these currents and voltages being transformed to a fixed two-phase reference frame associated with the stator, the apparatus comprising:

[0046] - A first flux estimator, used to estimate the flux of the motor in a fixed two-phase reference frame associated with the stator, comprising two electromotive force integrators for estimating the corresponding components of the flux along the direct and quadrature axes of the fixed two-phase reference frame associated with the stator, the electromotive forces being calculated based on the currents and voltages applied to each axis of the fixed two-phase reference frame associated with the stator.

[0047] - A torque estimator designed to estimate the motor torque based on the corresponding components of the flux estimated in the two axes of a fixed two-phase reference frame associated with the stator.

[0048] The device is characterized in that it comprises:

[0049] A second flux estimator is designed to receive the currents and voltages expressed in the fixed two-phase reference frame, which are transformed into a rotating two-phase reference frame associated with the rotor based on the rotor's position relative to the stator. The second flux estimator is designed to estimate the flux using the currents expressed in the rotating two-phase reference frame associated with the rotor. The second estimator includes an observer based on the stator currents and stator voltages expressed in the rotating two-phase reference frame associated with the rotor, which is designed to determine variables characterizing the magnetic uncertainty of the motor, said variables being used to correct the flux estimate of the second estimator.

[0050] - A detector designed to compare the torque estimated by the torque estimator with the setpoint torque, and when the difference between the estimated torque and the setpoint torque is greater than a predefined threshold, generate a signal that commands the initial conditions of the flux of the two integrators of the first flux estimator to be reset, the initial conditions being reset based on the flux estimate provided by the second flux estimator.

[0051] Advantageously, when the difference between the estimated torque and the setpoint torque is less than or equal to the threshold, the signal can disable the calculation of the second flux estimate by the second flux estimator. Attached Figure Description

[0052] Other features and advantages of the invention will become clearer from the following description, which is given as an illustrative rather than limiting example and with reference to the single accompanying drawing:

[0053] Figure 1 This is a functional block diagram of an apparatus for implementing the electromagnetic torque estimation method according to the present invention. Detailed Implementation

[0054] In the context of this specification, the case of permanent magnet three-phase synchronous motors used in hybrid and electric vehicle applications is therefore of particular interest.

[0055] In the first step, the current and voltage are measured for each of the three phases of the stator of the permanent magnet three-phase synchronous motor.

[0056] Next, the Parker transform is applied to the measured quantities involving these three phases to obtain the direct-axis current component i. d and cross-axis current component i q and the direct-axis voltage component v d and cross-axis voltage component v q .

[0057] Therefore, the three currents and three sinusoidal voltages of the three phases of the stator in a three-phase system can be transposed into a space in which the three sinusoidal current and voltage signals are expressed as two constant current and voltage signals (the components on the direct axis and the components on the quadrature axis q, respectively). For this purpose, the Parker reference frame is based on the reference frame associated with the rotor.

[0058] The torque estimator described in this context is based on the flux estimation of the electric motor, which is performed based on an improved voltage-flux model of the electric motor.

[0059] The estimation of the electromagnetic torque of the electric motor (expressed as) The following expression (1) provides the information:

[0060]

[0061] in:

[0062] -i α and i β These are the direct-axis and quadrature-axis components of the stator current in the Clarke domain, respectively, where α represents the direct axis of the Clarke domain and β represents the quadrature axis of the Clarke domain.

[0063] -λ α and λ β These are the direct-axis and quadrature-axis components of the total flux from the stator's perspective in the Clarke domain.

[0064] -p is the number of pole pairs of the electric motor.

[0065] The total flux from the stator of the permanent magnet synchronous motor can be estimated by calculating the integral of the electromotive force using the following set of equations (2):

[0066]

[0067] Among them, R S The resistance of the stator, v α and v β Let represent the direct-axis and quadrature-axis components of the stator voltage in the Clarke domain, respectively. and These represent the initial conditions for the flux based on the direct-axis and quadrature-axis components in the Clarke domain, respectively.

[0068] As will be explained in more detail below, in order to overcome the problem of estimating flux using the voltage model of the electric motor, this invention proposes to improve the model by combining it with a modified current model in order to reset the initial conditions of the integrator. and

[0069] More specifically, these initial flux conditions are determined using a current model and an appropriate observer designed to estimate the magnetic uncertainties of the model caused by magnetic saturation and demagnetization.

[0070] In practice, for low current levels, the relationship between flux and stator phase current is linear. However, at higher current levels, this linearity is no longer maintained, often leading to saturation operation. Furthermore, high-temperature operation of the motor causes changes in the flux of the permanent magnets (demagnetization). Therefore, due to magnetic saturation and demagnetization, the inductance of the motor and the magnetic flux of the permanent magnets change significantly in a predictable manner. Therefore, a modified flux-current model is proposed that takes these variations into account to improve the model's performance.

[0071] For this purpose, the unknown magnetic conditions of the motor in the current model will be considered. The unknown magnetic conditions considered are the linear inductance of the stator, the magnetic flux of the permanent magnet, and the saturation functions of the motor, which represent magnetic saturation and cross-saturation in the motor.

[0072] Considering these unknown magnetic conditions of the electric motor in the model yields the following system of equations in Park space:

[0073]

[0074] in and These are the components of the stator flux produced along the direct axis d and quadrature axis q in the Parker plane of the motor, respectively. d and v q These are the voltages applied to the direct axis d and quadrature axis q of the Parker plane of the motor, respectively. d and i q R represents the current flowing in the motor along the direct axis d and the quadrature axis q, respectively, in the Parker plane of the motor. S It is the equivalent resistance of the stator of the electric motor, L d0 and L q0 These are the linear inductances of the stator on each direct axis d and quadrature axis q in the Parker plane of the motor, respectively, φ. f It is the magnetic flux generated by the permanent magnets of the rotor, and ω e It is the angular velocity of the rotor.

[0075] f sat(1,2) It is a function that represents magnetic saturation and cross-saturation in an electric motor.

[0076] and:

[0077]

[0078] Among them, L d0m and L q0m These are the precise values ​​of the linearity coefficients between the flux and stator current on each direct axis d and quadrature axis q of the Parker plane of the electric motor, φ. fm It is the true value of the magnetic flux generated by the permanent magnets of the rotor, and f sat(1,2)m This is the corresponding true saturation function of the motor. Therefore, the parameters... Δφ f and Δf sat(1,2) This represents the difference between the true values ​​and reference values ​​of these parameters in the model proposed by equations (3) and (4).

[0079] Next, two variables, g1 and g2, are introduced into the current model, corresponding to the aforementioned unknown magnetic conditions. For this purpose, the stator current i is considered first.d and i q As a state variable, substituting it into equation (3) in equation (4) yields:

[0080]

[0081] in:

[0082]

[0083] Therefore, due to the change in inductance caused by magnetic saturation and the uncertainty in the permanent magnet flux, variables g1 and g2 group all the magnetic uncertainties of the motor together. In other words, the variables g1 and g2 to be observed can represent the magnetic behavior of the motor as a result of magnetic saturation. Introducing the parameter uncertainties described by variables g1 and g2 into the current model reflects the fact that the values ​​of the system's magnetic parameters (inductance and permanent magnet flux) may differ from those in the model due to motor saturation and demagnetization. As described below, observing these variables will allow correction for these uncertainties.

[0084] Therefore, an appropriate observer will be used to estimate these variables presented in the model. The basis is the aforementioned flux-current model of the electric motor, which takes into account the changes in the motor's magnetic parameters (inductance and permanent magnet flux) caused by magnetic saturation and demagnetization, and is based on observer theory, which includes the concepts of observability and state observers. Thus, when the state of the system cannot be measured, the observer can be designed and reconstructed based on the model of the system and measurements of other states. A state is understood to mean a set of physical values ​​defining the observed system. Here, a Kalman observer will be used to observe the magnetic uncertainties of the electric motor characterized by variables g1 and g2.

[0085] Therefore, variables g1 and g2 are considered as two new state variables. In this model, g1 and g2 are assumed to be piecewise constant functions, meaning their values ​​are constant functions over the interval. The system of equations modeling the electric motor can be reformulated in the general form of a nonlinear system:

[0086]

[0087] y = h(x)

[0088] in:

[0089]

[0090] and

[0091]

[0092] For the system modeled in this way, the observer can be formalized by the following equation:

[0093]

[0094] in It is the estimated state vector corresponding to the state vector:

[0095]

[0096] And K is the gain of the observer.

[0097] The choice of gain K, which is multiplied by the error term, allows for the adjustment of the observer. This gain is calculated using the Kalman algorithm.

[0098] The rotor's electric angular velocity is considered as input to the matrix system A. This allows for state observation. Therefore, a time-varying observer gain K is used, as follows:

[0099] K = P -1 C T R -1 (9)

[0100] The dynamic range of variable P is defined as follows:

[0101]

[0102] Matrix Q and R are positive definite symmetric weighted matrices.

[0103] Therefore, after estimating the two variables g1 and g2 using the observer, the total stator flux can be estimated based on the equations (4) and (7) defined above:

[0104]

[0105] As seen, the flux-voltage model of the electric motor is expressed by the above set of equations (2), based on which the flux of the electric motor can be estimated by calculating the integral of the electromotive force. Therefore, the estimation of the stator flux by the voltage model is based on the use of a resettable integrator, because, within the scope of the invention, as described above, the initial conditions of the integrator flux are determined by an independent observer, thereby enabling the flux to be estimated by applying the set of equations (11) based on a modified current model.

[0106] Therefore, a first estimator based on a voltage model is used to implement two pure integrators to estimate the flux in axes α and β corresponding to the two axes in a fixed two-phase reference frame, and in parallel, another flux estimator based on a current model in a rotating two-phase reference frame dq is used, which is modified with an observer to estimate the initial flux conditions for the two integrators.

[0107] Based on the flux estimation performed using a voltage model (which is thus improved by the use of an additional flux estimator based on a current model), the electromagnetic torque generated by the permanent magnet synchronous motor can then be calculated by applying expression (1).

[0108] To minimize computation time in real-time processing, it is proposed to stop observer-related processing when integrator initialization is no longer required. Therefore, a detector is proposed to define the initialization time. This detector will use the torque estimated by the voltage model (represented as...) ) and setpoint torque (expressed as T) * e The setpoint torque is compared to the torque to be provided by the vehicle's electric motor and given by the vehicle's control system, specifically based on the driver's depressing of the accelerator pedal. If the absolute value of the error between the setpoint torque and the estimated torque exceeds a predefined threshold ε, the observer is activated, and the two integrators of the flux estimator based on the voltage model are reset based on initial flux conditions determined by the flux-current model modified by the observer. Therefore, a Boolean operator for triggering the integrator reset is introduced into the voltage model, defined as follows:

[0109]

[0110] The observer is activated and deactivated when the operator "RES" is changed to "1" and "0" respectively.

[0111] It may be of interest to determine the percentage of time the observer is activated relative to the total operating time the electric motor is used to drive the vehicle. Therefore, a new monitoring variable is introduced, as follows:

[0112]

[0113] Where ∑t i It is the sum of the times the observer is activated, and t tot This is the total operating time. The monitored variable OT can vary from 0% to 100% based on unknown initial flux conditions at startup, magnet quality, motor saturation conditions, changes in setpoint torque, etc.

[0114] Measurement uncertainties and / or the nonlinearity of the inverter that supplies power to the motor can cause deviations in the calculated electromotive force. Therefore, to compensate for this deviation and prevent the estimated flux value from deviating over time, an average value of this deviation is proposed, as follows:

[0115]

[0116] Where T is the period of the current.

[0117] e αβ-偏移 This represents the offset of the calculated electromotive force. This value is intended to be subtracted from the online calculated electromotive force before it is used in the integrator of the flux estimator based on the flux-voltage model.

[0118] Figure 1 A block diagram is shown illustrating the method described above for estimating the electromagnetic torque of an electric motor.

[0119] First, collect the measured values ​​of current and voltage for each of the three phases of the stator of the permanent magnet three-phase synchronous motor.

[0120] By applying the Clarke transform, the projections of the measured stator electrical quantities involving these three phases onto two fixed axes (α, β) of a fixed two-phase (direct-axis and quadrature-axis) reference frame are obtained. Therefore, based on the voltage model, in Figure 1 The input of the device considers the current and voltage components i applied to the direct and quadrature axes of the Clarke domain, respectively. αβ and V αβ And provide it to the first flux estimator 10.

[0121] The first flux estimator 10 is initially based on the component i provided at the input. αβ and V αβ And the stator resistance Rs are used to determine the electromotive force e. αβ The direct axis component and the quadrature axis component on each of the two axes in the Clark domain.

[0122] Therefore, the first flux estimator 10 includes a multiplier 101 for each direct-axis and quadrature-axis component of the Clarke domain, which multiplies the stator resistance value Rs with the corresponding component i of the current of the considered axis. αβ The components are multiplied, and their outputs are provided to the corresponding adders 102. For each component considered, the corresponding adder 102 then subtracts the output of the multiplier 101 from its input receiving the corresponding stator voltage component, and outputs the electromotive force e of the considered axis. αβ The corresponding components are provided to the corresponding integrator 103. For each considered component, the corresponding integrator 103 calculates the integral of the electromotive force and provides at the output the direct-axis and quadrature-axis components of the total flux from the stator in the Clarke domain.

[0123] These direct-axis and quadrature-axis components of the total flux from the stator perspective in the Clarke domain The torque is provided to torque estimator 20, which calculates the electromagnetic torque provided by the motor by applying the above expression (1).

[0124] As explained above, the torque estimator outputs its value... The torque is provided to detector 30, which is designed to input the torque estimated by torque estimator 20 based on voltage model. With setpoint torque T * e The comparison is performed. Therefore, detector 30 includes adder 31, which takes the output provided by torque estimator 20 as input. Subtract this setpoint torque value T from the middle * e The output is provided as an absolute value to comparator 32. The comparator compares the absolute value of the error between the estimated torque and the setpoint torque with a predefined threshold ε and provides a Boolean operator RES at the output, which is designed to command the two integrators 103 of the first flux estimator 10 to reset when it changes to the value "1" (which indicates that the absolute value of the difference between the estimated torque and the setpoint torque is greater than the predefined threshold).

[0125] According to the invention, this reset of the two integrators 103 of the first flux estimator 10 is based on the activation of the second flux estimator 40, as explained above, which is based on a current model modified by an observer in the reference frame (d,q) to correct the estimated flux while taking into account the magnetic uncertainties of the motor model. Therefore, the signal RES is used to activate this second flux estimator 40 and preferably deactivate it when it changes to "1" and "0" respectively.

[0126] Parker converter circuit 41 uses the rotor position θ relative to the stator detected by the angular position sensor. e This is used to perform a rotation from reference frame (α,β) to reference frame (d,q). Therefore, after performing a Clarke transformation on the measured stator electrical quantities, there is a rotation of the axis, which will change the component i of reference frame (α,β). αβ and V αβ Transformed into continuous components along the two axes d (direct axis) and q (intersecting axis) of the Parker plane of the motor.

[0127] Therefore, the input of the second flux estimator 40 provides the current component and voltage component i along the two axes, the direct axis d and the quadrature axis q, respectively applied to the Parker plane of the motor. dq and v dq The second flux estimator is based on a current model modified with an observer.

[0128] The rotor's rotational speed ω is also provided at the input of this second flux estimator 40. e .

[0129] After estimating the two variables g1 and g2 using the observer as explained above, the second flux estimator 40 provides the direct-axis d-component and cross-axis q-component of the total flux in the Parker plane at the output by applying the above equation set (11).

[0130] These components are transformed using the inverse Parker transform 42, which is based on the rotor's position θ relative to the stator. e This is used to perform a rotation from reference frame (d,q) to reference frame (α,β). Therefore, the inverse Parker transform 42 provides the components of the total flux in the two-phase reference frames α and β at the output. These components are estimated based on the flux-current model of the electric motor, which is accompanied by an observer to account for the magnetic uncertainty of the flux-current model of the electric motor.

[0131] Therefore, when the signal RES of the two integrators 103 of the first estimator 10, which is reset by the second flux estimator 40 based on the flux-current model according to the flux-voltage model, is activated, these components are provided to the two integrators 103 to reset the initial conditions of the integrators. and

[0132] Figure 1 The device also includes a monitoring circuit 50, which is designed to measure the activation duration of the second estimator 40 with an observer using the aforementioned variable OT.

[0133] at last, Figure 1 The apparatus also describes a circuit 60 for estimating the deflection of the electromotive force, which provides at the output a value e determined in accordance with the principles described above. αβ-偏移 This offset value is provided to adder 102 so that the offset value is subtracted from the electromotive force calculated at the output of adder 102.

[0134] Therefore, even without knowing the initial magnetic conditions, the torque of the motor can be estimated in an accurate and robust manner by taking into account all the changes in the magnetic parameters during the operation of the motor.

Claims

1. A method for estimating the electromagnetic torque of a permanent-magnet three-phase synchronous motor of an electric or hybrid motor vehicle, the method comprising the steps of: - measuring the currents and the voltages in a fixed three-phase reference frame related to the stator of the motor on phases of the stator, - determining the currents and voltages (i αβ , v αβ ) in a fixed two-phase reference frame associated with the stator from the current and voltage measurements in the three-phase reference frame - determining the currents and voltages (i αβ , v αβ ) in a fixed two-phase reference frame associated with the stator from the current and voltage measurements in the three-phase reference frame - calculating a first flux estimate of the electric motor by calculating an integral of an electromotive force, the electromotive force being calculated based on the voltages and the currents expressed in a fixed two-phase reference frame associated with the stator, the first flux estimate, the electromotive force being calculated based on the voltages and the currents expressed in a fixed two-phase reference frame associated with the stator, - estimating the electromagnetic torque on the basis of the flux estimated by the first flux estimation the method being characterized in that it comprises the steps of: - determining currents and voltages in a rotating two-phase reference frame related to the rotor (i dq , v dq ) from the currents and voltages (i αβ , αβ ) expressed in the fixed two-phase reference frame and the position of the rotor of the electric motor relative to the stator (Θ e ), - calculating a second flux estimation of the electric motor, said second flux estimation being corrected by introducing a variable characterizing the magnetic uncertainty of the electric motor, said variable being determined by an observer based on the stator currents and the stator voltages (i dq , v dq ) expressed in a rotating two-phase reference frame related to the rotor, in parallel with the currents expressed in a rotating two-phase reference frame related to the rotor, ​ - comparing the torque estimate based on the flux estimate estimated by the first flux estimate with a setpoint torque (T * e ) and - resetting the initial conditions of the flux in compliance with the first flux estimate based on the flux estimated by the second flux estimate when the difference between the torque estimated based on the flux estimated by the first flux estimate and the setpoint torque exceeds a predefined threshold (e).

2. The method of claim 1, wherein, the method comprising the step of deactivating the calculation of the second flux estimate when the difference between the torque estimated based on the flux estimated by the first flux estimate and the setpoint torque is less than or equal to said threshold.

3. The method of claim 2, wherein, the method comprising the step of determining the percentage of the time of activation of the calculation of the second flux estimate relative to the time of operation of the motor.

4. The method of any one of claims 1-3, wherein, the method comprising the step of compensating for the offset of the calculated electromotive force before using it in the first flux estimate.

5. The method of claim 1, wherein, the second flux estimate of the flux of the motor comprising the steps of: - determining a flux-current model of the motor taking into account saturation functions of the motor so as to take into account in the model variations of the magnetic parameters of the motor caused by the magnetic saturation and demagnetization of the motor, - introducing into the model said variables characterizing the magnetic uncertainties of the motor, these magnetic uncertainties being inherent to the variations of the magnetic parameters of the motor, - estimating said variables in the model, this estimation being performed by said observer, - calculating the second flux estimate based on the estimated variables.

6. The method of claim 5, wherein, the magnetic parameters taken into account being the linear inductances of the stator on each axis of a rotating two-phase reference frame related to the rotor respectively and the magnetic flux produced by the permanent magnets of the rotor of the motor.

7. The method of claim 5 or 6, wherein, the saturation functions representing the magnetic saturation and cross saturation in the motor.

8. The method of any one of claims 1-3, wherein, the method comprising the step of adjusting the observer using a Kalman algorithm.

9. A device for estimating the electromagnetic torque of a permanent-magnet three-phase synchronous motor of an electric or hybrid motor vehicle, said device being designed to receive the currents and the voltages measured in a fixed three-phase reference frame related to the stator of the motor on phases of the stator of the motor, these currents and voltages being transformed into a fixed two-phase reference frame related to the stator, said device comprising: a first flux estimator (10) for estimating a flux of the electric motor in a fixed two-phase reference frame associated with the stator, the first flux estimator comprising two electromotive force integrators (103) for estimating respective components of the flux along two axes of the fixed two-phase reference frame associated with the stator The electromotive forces are calculated based on the currents and the voltages applied to each axis of the fixed two-phase reference frame associated with the stator (i αβ , v αβ ) a torque estimator (20) designed to estimate the torque of the electric motor on the basis of said respective components of the flux estimated in two axes of a fixed two-phase reference frame associated with the stator ​ the device being characterized in that it comprises: a second flux estimator (40) designed to receive the currents and the voltages (i αβ , αβ ) expressed in the fixed two-phase reference frame, transformed into a rotating two-phase reference frame related to the rotor as a function of the position (Θ e ) of the rotor relative to the stator, said second flux estimator (40) being designed to estimate the flux using the currents (i dq ) expressed in the rotating two-phase reference frame related to the rotor, said second flux estimator (40) comprising an observer based on the stator currents and the stator voltages (i dq , dq ) expressed in the rotating two-phase reference frame related to the rotor, the observer being designed to determine a variable characterizing the magnetic uncertainty of the electric motor, said variable being used to correct the flux estimate of said second flux estimator (40), and a detector (30) designed to compare the estimated torque with a setpoint torque (T estimated torque with a setpoint torque (T * e )) and to generate a signal (RES) capable of commanding the resetting of the initial conditions of the flux of the two integrators (103) of said first flux estimator (10) when the difference between the estimated torque and the setpoint torque is greater than a predefined threshold (ε), said initial conditions being reset on the basis of a second flux estimate provided by the second flux estimator (40).

10. The apparatus of claim 9, wherein, said signal (RES) being able to deactivate the calculation of the second flux estimate by the second flux estimator (40) when the difference between the estimated torque and the setpoint torque is less than or equal to said threshold.

Images