FR2825850A1 - Method of control of an alternating current electric motor, computes voltage vector angular offset arising from delay with slow processor performing control computations - Google Patents

Abstract

The method of control takes into account processor computation time. The tension applied to the motor is calculated for an angle \=H\=n which is equal to the angle \=H of the rotor at the start of the computation plus the phase offset corresponding to the displacement of the rotor during time taken for the processor to compute the required voltage.

Description

<Desc / Clms Page number 1>

 The present invention relates to a method for controlling an alternating current electric motor having p pairs of poles, supplied with sinusoidal current, in which the voltage to be applied to the motor is calculated as a function of the desired torque (reg) and of the angle @ of the rotor measured at the start of the calculation.

 Power steering systems for vehicles most often use electric motors, whether they are directly connected to the steering column or rack (EPAS type electric power steering) or whether they act on the column or the rack and pinion via a hydraulic system (electrohydraulic power steering type EHPS) It is important that the power steering system does not generate torque variations, as these are felt by the driver. This discomfort is particularly felt at low speed, for example in parking maneuvers or when driving at low speed. On the other hand, in an emergency, for example when a steering wheel thrust must be given, it is essential that the electric motor provides the assistance power necessary to assist the driver, the comfort of the driver being neglected.

 In some cases, brushless DC motors are used with a trapezoidal electromotive force. This type of motor is easy to control, but it often produces greater variations in torque than AC motors. In addition, they do not offer the possibility of control by defluxing.

 Another solution is to use AC motors, such as synchronous or asynchronous motors. In order to ensure dynamic control of the motor torque, these motors must be controlled by a directed flow control strategy. To be effective in a large speed domain, this strategy requires significant microcontroller resources or a floating point processor, which makes them relatively expensive. They cannot be used for large series because of their cost.

 To work around this problem, it would be desirable to use inexpensive microprocessors, but these microprocessors have a low computing power, which results in a low sampling rate and a significant computing time In this case, the applied voltage is not really sinusoidal and it has harmonics. When the electric motor is running at low speed, these deviations by

<Desc / Clms Page number 2>

 compared to the ideal sine curve are low and the motor torque supplied is practically constant. On the other hand, when the electric motor turns faster, for example in an emergency situation, the frequency of the current increases, but not the sampling rate and the calculation time remains constant. Consequently, the applied voltage is out of phase with respect to the advance taken by the rotor during the time necessary for acquisitions (setpoints, position and speed of the rotor, etc.) at the start of the calculation and at the calculation itself (grouped in the following the description under the term calculation time) and the control becomes unstable so that it can no longer regulate the torque.

 The objective of the invention is therefore to develop a method for controlling an AC electric motor which makes it possible to use inexpensive microprocessors with low sampling rate.

This objective is achieved by the method according to which the calculation time (Te) necessary for the control device to calculate the voltage to be applied to the motor in order to obtain the desired torque (trug) is taken into account in said calculation Taking into account the calculation time, it is possible to apply the desired voltage even with a high motor rotation speed. The motor is therefore able to provide the required assistance torque even in an emergency and the instability phenomena observed with traditional control methods do not occur
To take into account the calculation time of the microprocessor, the angular speed of the rotor (D) is measured, the phase shift angle (D'poTe) of the rotor between the start and the end of the calculation is calculated and the voltage to be applied is calculated for the angle @ + Qp Tc.

 The invention also relates to a device for implementing the method and to power steering systems implementing this method, in particular electric power steering of the EPAS type and electro-hydraulic power steering of the EHPS type.

 Although the method can be applied to the control of asynchronous motors, it is preferable to use synchronous motors with permanent magnets. Indeed, these engines are smaller and have better performance. They therefore help to save fuel.

<Desc / Clms Page number 3>

 It is in accordance with the invention that the microprocessor controlling the electric motor has a low sampling rate, preferably of the order of 0.5 to 1 ms.

The invention will be presented below using the following figures:
Figure 1: Example of control scheme
Figure 2: schematic representation of the phase difference between the beginning and the end of the calculation.

 The control of the motor is done in a conventional manner by a strategy of control by oriented flow, also known by the name of vector control.

p étant le nombre de paires de pôles du moteur et OR le flux du rotor. The two currents I, q and Id are obtained by transforming the motor currents in a coordinate system defined by the angle 0 of the rotor. This transformation of coordinates leads to an orientation of the flow of the motor on the axis d of the coordinate system of the flow dq. The engine torque is then expressed as follows:

p being the number of pairs of poles of the motor and OR the flux of the rotor.

d'assistance souhaité Treji de la vitesse du moteur Q et de la tension de la batterie Vba,. Le défluxage indirect basé sur la vitesse Q est obtenu en introduisant un composant de courant (Isdref) sur l'axe d. La zone de défluxage, et par conséquent la vitesse maximale à laquelle le couple d'assistance peut être produit dépend de la tension de la batterie, celleci allant généralement de 9 à 16 V. The reference currents Isqref and Isdref are calculated as a function of the torque

assistance required Treji of motor speed Q and battery voltage Vba ,. Indirect defluxing based on speed Q is obtained by introducing a current component (Isdref) on the d axis. The defluxing area, and therefore the maximum speed at which the assistance torque can be produced, depends on the battery voltage, this generally ranging from 9 to 16 V.

 The two current components Isq and Isd are regulated by regulators Pl.

The decoupling between the stator currents and the stator voltages is achieved by compensation of the electromotive force signals FEMd and FEMq.

 The transformation of the components of flow-oriented voltage Vrefq and Vrefd into three-phase stator current is carried out for the corrected rotor angle ev and not the angle 0 of the rotor obtained at the start of the calculation.

<Desc / Clms Page number 4>

 In fact, in inexpensive microprocessors, the computation time represents approximately 30% of the load of the microprocessor, or approximately 0.3 ms for microprocessors with a sampling rate of 1 ms. This significant calculation time results in a phase shift between the position of the rotor at the start of the calculation (@) and its position at the end of the calculation (ev). This offset can be neglected as long as the engine speed is low. On the other hand, when the motor rotates more quickly, this phase shift becomes significant and the control becomes unstable so that it can no longer regulate and the desired motor torque is not supplied.

commande. On utilisera donc dans la suite de la description les termes de degrés (0) mécaniques pour noter un angle parcouru physiquement par le rotor du moteur et de degrés (") électriques pour un angle de phase relatif aux signaux électriques de commande, comme par exemple l'angle du vecteur tension qui sera détaillé avec la figure 2. En l'occurence, pour un moteur ayant deux paires de pôles, une rotation d'un degré mécanique du rotor correspond à une rotation de deux degrés électriques. Dans la suite de la description, sauf mention contraire, les angles cités correspondront à des degrés électriques. It will be recalled that for a synchronous motor comprising p pairs of poles, a fixed relation links the speed of rotation of the motor to the frequency of the electrical signals of

ordered. The terms of mechanical degrees (0) will therefore be used in the following description to denote an angle physically traversed by the motor rotor and electrical degrees (") for a phase angle relating to the electrical control signals, such as for example the angle of the voltage vector which will be detailed with FIG. 2. In this case, for a motor having two pairs of poles, a rotation of one mechanical degree of the rotor corresponds to a rotation of two electrical degrees. the description, unless otherwise stated, the angles quoted will correspond to electrical degrees.

 Taking the example of a four-pole motor rotating at 3000 revolutions 1 minute, the frequency of the stator current is 100 Hz. Consequently, the rotor will have rotated by 5.4 0 mechanical by the time that the microprocessor calculates the voltage to be applied to obtain the desired torque. It is therefore necessary to correct the voltage of 10.8 "electrical to take account of the rotation of the rotor. This principle is shown in Figure 2.

< 9p = < 9 + zf < 9 = < 9 + -y [3] In practice, the computation time Tc of the microprocessor used is known. The rotation speed of the motor Q and the angle @ of the rotor are traditionally measured at the start of the calculation. The electrical phase shift Lie corresponding to the advancement of the rotor between the start and the end of the calculation is calculated according to the following formula: Li e = Q'poTe [2] and the angle of the rotor ev at the end of the calculation is therefore the next :

<9p = <9 + zf <9 = <9 + -y [3]

<Desc / Clms Page number 5>

The voltage to be applied to the motor is then calculated not for e but for ev.

 This control strategy taking account of the calculation time makes it possible to use alternating current motors with microprocessors of low calculation power.

The sampling rate of these microprocessors is sufficient to obtain an almost constant torque for normal low speed uses of the electric motor, such as parking maneuvers or steering while driving. By correcting the angle in the Park transformation for calculating the voltages to be applied, it is possible to use this command even for high rotational speeds, which makes it possible to ensure an efficient torque even in an emergency ( eg steering wheel kick). In these emergency situations, the torque is no longer constant due to the increase in the frequency of the current, but in the case of a temporary emergency situation, the comfort of driving can be neglected in favor of safety, that is to say the efficiency of the assistance provided to the driver.

 This process makes it possible to use inexpensive processors while providing good driving comfort at least for normal conditions of use (parking maneuver and driving direction). The use of permanent magnet synchronous motors, which are more compact and more efficient than asynchronous motors, also allows additional fuel savings. Furthermore, the implementation of the method does not require additional electronic resources.

 Although the example cited shows the use of the method for a power steering system, this method and its implementation device can be used for other applications which require economical implementation and in which the variation in torque at high speed does not play a major role.

Claims (9)

 CLAIMS 1. Method for controlling an alternating current electric motor having p pairs of poles, supplied with sinusoidal current, in which the voltage to be applied to the motor is calculated as a function of the desired torque (Trer) and of the angle 0 of the rotor measured at the start of the calculation, characterized in that the calculation time (Tc) necessary for the control device to calculate the voltage to be applied to the motor in order to obtain the desired torque (Trer) is taken into account in said calculation.  2. Method according to claim 1, characterized in that the angular speed of the rotor (Q) is measured, the electrical phase shift angle (Qp TC) of the rotor between the start and the end of the calculation is calculated and the voltage to be applied is calculated for the angle @ + Qp Tc 3. Device for implementing the method according to claim 1, characterized in that it comprises means for calculating the phase shift angle 4. Power steering system adapted to the implementation of the method according to claim 1, comprising at least one electric AC motor, means for measuring the position and the angular speed of the rotor of said motor, and a device for control adapted to calculate a voltage to be applied to the motor as a function of the desired torque (trug) and of the angle e of the rotor, characterized in that said control device is programmed to take into account its calculation time (Te) in the calculation of said voltage.  5. Power steering system according to claim 4, characterized in that the alternating current electric motor is a synchronous motor with permanent magnets.  6 Power steering system according to claim 4, characterized in that the electric motor acts directly on the steering column (electric power steering).  7. Power steering system according to claim 4, characterized in that the electric motor acts on the steering column by means of a hydraulic system (electro-hydraulic power steering). <Desc / Clms Page number 7>  8. Power steering system according to claim 4, characterized in that the microprocessor controlling the electric motor has a low sampling rate.  9. Power steering system according to claim 8, characterized in that the sampling rate is of the order of 1 ms.