WO2008043939A2

WO2008043939A2 - Method for identifying the malfunction of a vehicle actuator, vehicle actuator and vehicle including actuators according to the present method

Info

Publication number: WO2008043939A2
Application number: PCT/FR2007/052080
Authority: WO
Inventors: Xavier Claeys; Lionel Cordesses; Marco Marsilia
Original assignee: Renault S.A.S.
Priority date: 2006-10-11
Filing date: 2007-10-04
Publication date: 2008-04-17
Also published as: FR2907089B1; WO2008043939A3; FR2907089A1

Abstract

The invention relates to a method for identifying a malfunction of an electric motor of an electric power system, in particular an electric power steering for a vehicle. The invention also relates to an electric power steering for vehicle implementing this method. The electric power steering co-operates with a malfunction identification circuit integrating a state representation model and with residue processing means for deducing the identification of a malfunction in the steering assistance electric motor.

Description

The present invention relates to a method for identifying the malfunction of a machine for a vehicle actuator. , an electric motor electric power steering member, or as a drive actuator for a motor vehicle. The invention also relates to a vehicle incorporating such.

In the state of the art, vehicle directions are known which make it possible to assist the steering effort of the driver with the aid of an electric motor. Such a system is designated by the term "electric power steering". Moreover, the electric motor, alone or in combination with a heat engine, provides more and more vehicle driving functions.

In an assisted steering system, an electric motor decouples the forces of a user of a mechanical load constituted by the steering mechanism itself and acting on the steering wheels.

Particularly, when the concept is applied to a vehicle steering, an electric motor is controlled by the steering wheel maneuver so that steering wheel steering can drive the vehicle to the right or to the left.

However, operational safety issues are of crucial importance for electric power steering. In the state of the art, it has already been proposed to carry out a method of detecting malfunction of a power steering system for a vehicle. Such a state of the art is notably known from document FR-A-2.849.420 by the same applicant in which a method and a system for diagnosing malfunction for a decoupled vehicle are proposed. In this document, a malfunction detection method is particularly described in which a position of the steering wheel is measured to give a steering instruction to the vehicle, and a second magnitude related to the operation of the steering wheel is also measured as the torque exerted on the steering wheel. steering wheel so as to be able to diagnose a steering wheel-type malfunction.

However, such a method does not solve all malfunction situations, particularly in the case of a power assisted steering in which the electric motor can suffer many causes of malfunction.

In addition, the state of the art does not provide means for resisting disturbances caused by random noise or other causes related in particular to the deformation of the vehicle frame, or reactions between the steering wheels and the road on which the vehicle is moving.

In US-A-6,640,167, a fault detection method is described which relies on the development of a residue for detecting an error. However, the method has shortcomings because it requires a large number of redundancies to not make false directions. Several successive tests are carried out to check the defects involving the redundancies but no guarantee of the insulation of the failure is acquired.

With the difference, the document FR-A-2.849.420 already cited, describes a way to determine a malfunction on a direction by measurement analysis and consistency tests. This This method has the advantage of being less dependent on the parameters of the system, but risks providing false directions under the influence of external disturbances.

Document US-A-6,729,432 describes a method for detecting defects integrated in a vehicle. This method has several disadvantages. In particular, the method does not make it possible to carry out detectability tests and thus to better choose the measurements that are useful for detection. Moreover, the residue used is analyzed only by thresholding. On a highly noisy system it will not be possible to obtain a convincing result.

In a vehicle drive system, one or more electric motors, sometimes reversibly working in brakes, for example in a so-called regenerative braking phase, exchange mechanical power with the wheels of the vehicle.

As for the safety of a steering system, the safety of a vehicle drive system is essential for driving safety and this safety would be jeopardized by any malfunction of the electric machine that cooperates.

It is an advantage of the method of the invention to provide an analysis of the dynamic system constituted by the vehicle to identify the most appropriate measures to isolate the defects of the electric motor. It does not claim to strongly limit the number of measurements, but at least to choose a sufficient number of measures in agreement with the failures that one wishes to identify.

It is another advantage of the method of the invention, by virtue of a particular filtering of the residues, to provide a more efficient analysis of residues and thus to be able to detect even small operating defects. In particular, the invention aims to provide a system and a method for detecting dysfunction less sensitive to false detections.

Indeed, the present invention relates to a method of identifying malfunction of an electric motor.

According to the invention, the method comprises a preliminary step for generating a dynamic state model of the vehicle, then a step for identifying the faults acting on the electric steering motor. The invention also relates to an electric vehicle actuator, such as an electric power steering of the kind comprising at least one electric steering assist motor decoupled from a steering mechanism. The invention is characterized in that the electric actuator cooperates with a malfunction identification circuit incorporating a state representation model and a residue treatment means to deduce the identification of a malfunction of the electric motor of the electric actuator.

According to one aspect of the invention the direction, to implement the malfunction identification method of the electric motor of the actuator comprises:

a block for detecting sensors and for shaping the signals coming from the sensors associated with the operation of an electric actuator such as an electric power steering; a block in which the signal originating from the sensors is processed so as to produce a vector signal for processing sensors;

a block which executes an isolator filtering on the basis of the vector signal for processing the sensors and supplies at its output a filtered signal isolated from the aforementioned vector signal;

a residue treatment block according to a method of functional analysis in the state space also derived from a modeling the dynamics of the electric actuator as an electric power steering system;

a block applying a decision algorithm whose output signal is a vector decision signal which is transmitted to the input of the last block; and

an output block in which the decision signal is processed so as to provide signaling of a malfunction or defect.

According to another aspect of the invention, the detector filtering circuit comprises means for implementing a state representation model.

According to another aspect of the invention, the means for implementing a state representation model comprises a reduction means of the complete model taking into account the electrical dynamics of the motor as well as the movement of the equivalent weight of the motor. assembly of the electric actuator as that constituted around the steering mechanism composed of the electric assist motor, the steering rack and the steering wheels. According to another aspect of the invention, the detector filter comprises a discrete linear filter making it possible to detect the defects of the electric motor.

According to another aspect of the invention, the discrete linear filter comprises a means for calculating disturbance-related residues intended to identify the malfunctions on the basis of the observable states of the state representation model.

According to another aspect of the invention, the discrete linear filter comprises means for decoupling the residues of each identifiable dysfunction. According to another aspect of the invention, the decoupling means comprises a discrete filter synthesis circuit based on a determined set of algebraic constraints representative of the malfunctions. According to another aspect of the invention, the discrete synthesis circuit comprises means for discretizing the input signals described in the reduced state representation on the basis of a sampling clock k. According to another aspect of the invention, the model reduced by state representation is determined by the data of matrices with constant coefficients or time evolution (A _r , B _r , F _r , C _r ).

According to another aspect of the invention, the malfunction identification circuit comprises sampling circuits controlled by the clock k, the discrete filter establishing the discretized reduced model determined by the data of the matrices with constant coefficients or with temporal evolution. sampled according to k (A _d , B _d , F _d , C _d ) used in the discretized reduced model. According to another aspect of the invention, the means for forming the residues concerned for the identification of the malfunction of the electric motor comprises an operator who executes the relation:

D = C _d χ F _d and includes a means for implementing algebraic constraints representative of the malfunctions.

According to another aspect of the invention, the detector filter circuit comprises an operator that executes the relation: q _k = U x (γ _k - C _d x X _k ) which determines the k-th residue for the identification of the malfunction on the implementation of a means for detecting a time delay equal to its detectability index.

According to another aspect of the invention, the malfunction identification circuit includes means for setting the time sampling step k so that defects and associated residues can be discernible with a given detectability index. According to another aspect of the invention, the residue processing block comprises a digital filter which comprises means for increasing the signal-to-noise ratio so as to separate the disturbances from the measurement noises. According to another aspect of the invention, the digital residue treatment filter comprises:

means for attenuating the amplitude of the unwanted signal by a certain factor;

means for maintaining the slope (first order derivative) of the disturbance.

According to another aspect of the invention, the digital filter comprises two circuits producing a loopback integrator with negative feedback.

According to another aspect of the invention, the digital filter comprises an order integrating circuit "1" whose input u is connected to a first output of an amplifier circuit whose second output makes it possible to produce the filtered residue signal qf, in that the amplifier comprises a first input which receives the residual signal q from the block and a second input which is connected to the output iu of the order integrator "1".

According to another aspect of the invention, the integrator comprises a circuit executing the Euler method which at the iteration P + 1 uses the preceding sample p in a relation of the form: / w [p + 1] = iu [p] + Te ^* u [k] for a given integration time Te.

According to another aspect of the invention, the loop gain amplifier comprises an operator performing the function F () defined by: F (x) = Kj; if aμi <| x | > a-, with i> 0

For x = q - iu According to another aspect of the invention, the amplifier comprises means for eliminating the values Kj which do not satisfy the relation Ki <Kj for i <j.

According to another aspect of the invention, the amplifier comprises an output circuit which executes the linear operation: qf = G97 * lu + Offset in which G97 is a given gain and Offset a predetermined value, which may be zero.

According to another aspect of the invention, the amplification circuit co-operates with a circuit for adapting the dynamics of the filtered signal qf according to the vehicle.

According to another aspect of the invention, the output of the digital filter of the residue is connected to a comparator at a predetermined threshold to produce a Boolean signal of presence or absence of detected fault.

According to another aspect of the invention, the threshold value is produced by a threshold adapter circuit according to the desired sensitivity of the detectability of the malfunctions.

According to another aspect of the invention, the output of the digital filter of the residue is connected to a circuit producing a signal proportional to the intensity of the identified defect.

According to another aspect of the invention, the malfunction identification circuit of the invention comprises a malfunction detection block 9, connected in particular to the output of the residue filtering circuit which comprises a comparator at a predetermined threshold, of so that the output of the threshold comparator produces a Boolean signal of presence or absence of detected malfunction.

According to another aspect of the invention, the fault signaling block comprises means for displaying the presence of a fault on the electric torque of the electric steering assist motor and means for displaying the intensity or level of the fault.

Other advantages and features of the present invention will be better understood with the help of the description and the figures among which:

FIG. 1 is a block diagram of a circuit implementing the method of the invention;

- Figure 2 is a diagram showing a vehicle with the indication of various essential parameters to describe both the vehicle dynamics and the influence of external disturbances;

FIGS. 3A and 3B are diagrams representing the modeling elements of the interaction between the train of the steered wheels and the electric steering servo motor; FIGS. 4A to 4C show the graphs of test disturbance applications that can be used in the malfunction detection method of the invention;

FIGS. 5A to 5D show four signal graphs processed during a fault identification on an output parameter produced according to the method of the invention;

FIG. 6 represents a block diagram of a detector filter used in the circuit of a power assisted steering according to the invention; and FIG. 7 represents another nonlinear filter used in a particular embodiment of the circuit of a power assisted electrical direction for implementing the method of the invention.

The method of the invention comprises a preliminary step that is performed during the design of the vehicle and a step of identifying a malfunction of the electric motor during normal operation of the vehicle. In the remainder of the description, it is attached to describe an electric actuator constituted by an electric power steering system. The the same means, possibly transposed according to its own characteristics, are implemented for other electric actuators, such as one or more electric motors for driving the vehicle and they will not be described here.

To implement the method of the invention, the electric power steering which comprises a steering mechanism based on an electric motor which receives a torque setpoint to perform the steering assistance indicated by the driver acting for example on a steering wheel. steering, includes a circuit for signaling the detection of such a defect and which implements the first and second main step of the method of the invention. For this purpose, the circuit for signaling a malfunction of the electric steering motor cooperates with a computer in which the descriptive model of the evolutions is executed as a function of the demands of the driver of the physical states of the modeled system and which is established on the basis of the preliminary step of the process of the invention. The signaling circuit also cooperates with a fault identification circuit of the electric motor of the steering system in which the operation step of the method of the invention is performed.

This results in an alarm signal of a fault or malfunction or fault of the electric motor. The alarm signal is operated by the malfunction signaling circuit to enable either a driver safety action or the triggering of a turning of the steering system in a mode in which assistance is disabled. FIG. 1 shows a block diagram of the electrical steering motor electric motor malfunction identification circuit in which the operating step of the method of the invention is carried out and intended to be mounted. in a control system of an electric vehicle power steering.

In the method of the invention, the operating step, which is a step for identifying the faults acting on the electric motor of the steering system, comprises: a first step for detecting the instantaneous values of the system state vector represented as well as the engine torque setpoint;

a second step of filtering and shaping the signals representative of the values and setpoints detected in the first step;

a third step for performing a detection filtering operation to produce a description of the evolution of the measurable states of the system represented and the influence of the disturbances applied in the form of residues;

a fourth residue treatment step obtained in the third step for increasing the signal-to-noise ratio on at least one residue;

a fifth step of detecting the defect on the basis of the detection of at least one treated residue during the fourth step; and

a sixth step for exploiting the detection of at least one fault of the fifth step, in particular by signaling the detected fault as the display of an alarm. The malfunction identification circuit represented in FIG. 1 and which implements the operation step of the method comprises a series of six functional blocks which are respectively:

a block 1 for detecting sensors and for shaping the signals coming from the sensors associated with the operation of an electric power steering and which will be described later; a block 3 in which the signal coming from the sensors is processed so as to produce a vector signal for processing sensors;

a block 5 which executes an isolator filtering on the basis of the vector signal for processing the sensors and supplies at its output a filtered signal isolated from the aforementioned vector signal;

a residue treatment block 7 according to a method of functional analysis in the state space also derived from a modeling of the dynamics of the electric power steering system;

a block 9 applying a decision algorithm whose output signal is a vector decision signal which is transmitted to the input of the last block; and

an output block 1 1 in which the decision signal is processed so as to provide signaling of a malfunction or defect.

In Figure 2, there is shown a modeled vehicle 13 for exposing the control parameters used in the method of the invention. The model vehicle seen from above comprises four wheels 15 - 18 including two steering wheels 15 and 16 before and two wheels 17 and 18 rear, symmetrical on either side of a central axis 20 of the vehicle. In a one-dimensional modeling part, the modeled vehicle has a rear ground connection 26 and a ground link 24 before steering. The longitudinal, common steering angle of the front wheels 15 and 16 or of the ground connection 24 before steering is measured by an angle noted θ.

On the central axis of symmetry 20 of the vehicle, there is the center of inertia G on which can be associated a torsor comprising the velocity vector V of the vehicle and its transverse acceleration γ _t . Similarly, there is shown with an arrow denoted ψ 'the angular speed of rotation of the vehicle 13 measured about the normal axis of the vehicle traffic plane, which is called yaw rate. In particular, the relative differences between the steering angles θ and drift angle β on the one hand and the yaw rate and the integral of the transverse acceleration on the other hand make it possible to model the effects of the external disturbances undergone by the vehicle and make it possible to with the other parameters to model the dynamic behavior of the vehicle under the effect of the steering mechanism.

According to the method of the invention, its preliminary step comprises a step for establishing a state representation of the complete dynamic system that constitutes the vehicle in a driving situation with an electric power steering equipped with at least one electric motor to provide a torque assistance to the steering mechanism. The state representation that takes up the fundamental relationship of the dynamics applied to the vehicle

identifies the second time derivative - E 'of a state vector E, dt of the first time derivative E', of the system represented, to the sum of the instantaneous state of the system represented, of the electric state of the electric motor and of the instantaneous state of the disturbances applied to the represented system. In a particular embodiment of the invention, the dynamic behavior of the vehicle has been modeled so that it can be referenced by a matrix system of dynamic equations which is written in the form:

The vector E comprises six components under their direct form or their time derivatives. The transpose of the vector E is written in matrix line form:

[Ψ 'β & θ i r] whose six components are respectively from left to right:

• the yaw rate already explained; • the angle of drift already explained;

• the angular rate of longitudinal deflection

• the longitudinal steering angle already explained;

• the electric current flowing through the electric machine of the steering mechanism;

• the time derivative of the aforementioned electric current.

We will now describe how, in the method of the invention, we establish the instantaneous state of the system shown which is the term A x E of the above relationship. The matrix A is thus a square matrix of dimensions 6 (six) composed in four blocks according to the following form:

A = AA ₂ ^'

A -21A ₂ in which the four blocks are matrices of coefficients independent of the components of the aforementioned vector E. It has been defined in the invention:

D ₁ L ² D ₂ U ₂ D ₂ L ₂ - D ₁ L ₁

Ai = which connects the parameters y / 'and β D ₂ L ₂ - D ₁ L ₁

- 1 +

MV ^L MV with their derivatives

A ₉ = θ

and the current and its derivative of the electric motor of the steering mechanism;

A ₂₁ = 0 0 which links the derivatives θ \ θ and the currents / and / '0 0 0 0 to the parameters of ψ' and β; At ₁₁ θ and the

currents and / or their derivatives.

The matrix B of the second term of the matrix equation of the dynamic behavior of the vehicle equipped with the controlled steering mechanism according to the method of the invention is composed of five null coefficients and its sixth coefficient is written in the form:

1

B _Λ = - which is associated with the derivative of the current / 'in the motor ak. electric steering mechanism.

The vector of perturbations δ which is in the product of the third term of the right-hand side of the matrix equation of the dynamic state of the vehicle system, is a column vector with three components and it will be noted: δ M δ = δ _τ δ , C, in which the components are the instantaneous measurements or estimates, respectively from top to bottom, of the perturbation:

• moment type on the chassis (when applying the relation of the dynamics in rotation with a moment);

• type of force on the chassis (when applying the fundamental relationship of the dynamics in translation with a slider); and

• of electric current type or, preferably of electric torque type, considered as the product of the speed of rotation by the current), produced by the electric motor of the steering mechanism.

The third matrix C of the model is determined by: 0 ψ 1 0 0

MV

C = 0 0 0 0 0 0 0 0 0 1 0 0

The different coefficients of the four blocks of the matrix A and the two other matrices B and C of the state representation have the following meanings: V: speed of the vehicle; M: total mass of the vehicle;

I _Ψ : moment of inertia of the vehicle with respect to the axis orthogonal to the plane of the movement of the vehicle; l_i: distance from the center of inertia G of the vehicle to the front axle; L ₂ : from the center of inertia G of the vehicle to the rear axle; Di: rigidity coefficient of the forward drift; D ₂ : rigidity coefficient of the rear drift; k _c : electric torque constant of the electric motor of the steering mechanism;

J _eq : moment of inertia of the steering mechanism with the electric motor; λ: reduction ratio of the gearbox which mechanically couples the electric motor of the steering mechanism to the steering rack; a: dimensionless coefficient characteristic of the inductive behavior of the electric motor; b: dimensionless coefficient characteristic of the resistive behavior of the electric motor.

In order to establish the status representation of the dynamic system comprising the vehicle and its electric power steering, it is therefore necessary to establish each of the parameters indicated above or determine its state by modeling the vehicle at each possible value of its dynamic behavior.

FIGS. 3A and 3B show elements making it possible to model the vehicle during a steering operation using power assisted steering with an electric motor.

In FIG. 3A, a control member 29 produces a setpoint parameter constituted by the desired wheel angle which is transmitted to a control member 28 which represents a position control of the wheels. Such position control is achieved by the electric power steering system and will not be described further below.

This servocontrol device produces an output signal in the form of a control for example of the power supply voltage of the electric motor coupled by its gearbox and, if appropriate, a clutch to the steering mechanism voltage which is applied to this electric motor. steering assistance 30 which has a moment of inertia Jm.

A gearbox 31 applies a rotational speed ω and a determined torque C _e to produce on the steering mechanism 32, constituted by the steering axle, a deflection of which the angle and the angular velocity respectively θ and θ 'are detected by means of sensors 34 whose signals are returned to the servo member 28 which executes the control law K determined independently of the method of the invention. The command K is continued as long as the wheel angle setpoint is not adapted to the measurement.

FIG. 3b shows a detail of the steering mechanism which comprises a mechanism support 36 on which is mounted the steering assist electric motor 39 whose rotor shaft connects a moment of inertia of the rotor Jm to a gearbox, for example of worm-wheel type 40. The gearbox 40 is characterized by a reduction ratio λ as a ratio of the speeds of rotation applied to its input and produced at its output shaft. The output shaft of the gearbox 40, generally associated with the worm gear, is coupled in known manner to the conventional steering rack 32 of the electric power steering on which rack, the steering wheel (not shown applies the steering forces of the The rack 42 is characterized by a mass Mcrem and applies its forces including those of the driver and / or those of the electric steering assist motor 38 to a wheel pivot 44 so as to apply a torque or steering effort on the steering wheel shown 46 which has a moment of inertia Jr with respect to the pivot 44.

FIGS. 4A to 4C show the specification of three disturbances applied to the dynamical system shown and that, according to the method of the invention, the electrical motor malfunction identification circuits of the steering mechanism are capable of distinguishing between different situations. breakdowns.

In Figure 4A, there is shown a disturbance on the electric torque for example applied by a shock on the steering axle and taken into account in the third term of the state representation described above. At the date t0, a disturbance step of electrical torque Ce is applied to the electric steering assistance motor to a level A = 0.08 N.m in the example. The disturbance then falls after a predetermined duration of the order of 10 seconds with a slow decrease. Then, a negative -B amplitude perturbation at -0.06 N. m is applied to the date t4 with a duration of the order of 10 seconds.

FIG. 4B shows a stress-type perturbation produced for example by a lateral force on the chassis of the vehicle represented by a gust of wind applied on the date t2 and of a determined amplitude C, of 1000 N in an example test. The duration of application is of the order of 5 seconds. FIG. 4C shows a moment-type perturbation produced, for example, by an asymmetric variation in the adhesion of the wheels of the vehicle applied to the date t3 and which first shows a rise to a determined value D of 900 N in the example of test, then quickly negative of a value -E of the order of -400 N, followed by a return of slope C towards the neutrality of the perturbation.

We will now detail a particular embodiment of the six blocks of the electric motor malfunction identification circuit of the steering system in which is performed the operating step of the method of the invention.

In input block 1, the measurements which are used for the implementation of the state representation explained above are executed. For these measures, we use:

an electric current sensor i of the electric steering assistance motor;

a sensor for angular position and angular speed of the electric motor; and a detector of the electric torque setpoint C _e produced by the circuit 28 of the electric power steering.

In a particular embodiment, the angular position and angular velocity sensor of the aforementioned electric motor is replaced by an angular position sensor and steering wheel angular speed sensor and a calculation circuit implements a determined function f () which connects the steering angle q with the rotation angle of the motor θm in a multiplicative relation of the form: θ = f (λ; L) x θ _m with the notations previously exposed. The function of converting the steering angle into an instantaneous angle of rotation of the engine is a geometric function that takes into account the mechanical configuration of the components interposed between the steering axle and the electric motor with the wheel pivot, rack and reducer including.

The second block 3 executes a determined function of filtering the signals of the sensors of the block 1, to eliminate the possible noise of measurement and to put the signals in the desired format to execute the method of identifying the dysfunction of the invention.

The third block 5 performs a function of filtering detector in the sense of the identification of malfunctions. In a particular embodiment, the detector filter 5 implements a reduced state representation model relative to the complete model explained above. The model reduction carried out on the basis of the complete model described above consists in executing the definition of the electric dynamics of the engine as well as the movement of the equivalent mass of the entire steering mechanism composed of the electric assistance motor, the steering rack and steering wheels. A reduced vector of state X _r of the above-mentioned vector E is used, which comprises, with the preceding notations, the following four parameters: X ₁ . = [θ 'θ if] and Y _r = [θ' θ i] which is a vector extracted from the reduced state vector X _r .

The reduced state model is then written: X ' _r = A _r x X _r + B _r x C * + F _r x A

Y _r = C _r x X _r in which the matrix A _r of the first term is the block A ₂₂ of the complete model, B _r is the lower block of the matrix B of the complete model and F _r is a matrix directly obtained on the basis of of the matrix F and which is written:

1 0

0 0

F = 0 0 1 0 ak _c _ The matrix F _r represents the influence of the defects Δ on the states of the continuous reduced system with, respectively, for its first column the influence of the defects on the dynamics of the vehicle and for the second column the defects on the electrical torque. The matrix C _r simply makes it possible to extract the vector Y _r from the reduced vector of state X _r with:

1 0 0 0

C = 0 1 0 0 0 0 1 0

The vector Δ, used in the third term of the reduced state representation, makes it possible to synthesize the disturbances taken into account so as to separate the identified disturbances from identifiable failures according to the invention, even with a reduced state representation:

ψ ^ ψ '+ ^ β

Δ = ^• VJ eq in which we recognize the effect of the first δ. line of the block A ₂ i of the first term of the complete representation of state and the term of disturbance on the electric torque.

On this basis, the detector filter of the block 5 comprises a discrete linear filter for detecting the defects of the electric motor. By using the theory of observable states, such a discrete filter makes it possible to calculate the residues associated with the disturbances in which the method of the invention makes it possible to identify the malfunctions. For this purpose, the discrete linear filter of the invention comprises a means for decoupling the residues of each identifiable dysfunction. In a particular embodiment, the decoupling means comprises a discrete filter synthesis circuit on the basis of a determined set of algebraic constraints representative of the malfunctions.

The discrete synthesis circuit comprises a means for discretizing the input signals described in the representation reduced state based on a sampling clock k. The discretized reduced model is then written:

X ' _{k + ι} = A _d x X _r + B _d xu _k + K _k (Y _k - Y _k )

Ϋ _k = C _d x X _k The model by reduced representation of states is determined by the data matrices with constant coefficients or temporal evolution (A _r, B _r _n F _n C) described above. By sampling circuits controlled by the above-mentioned clock k, the discrete filter establishes the discretized reduced model determined by the data of the matrices with constant coefficients or with sampled time evolution according to k (A _d , B _d , F _d , C _d ) used in the discretized reduced model defined above.

In a particular embodiment, the means for forming the residues concerned for the identification of the malfunction of the electric motor comprises an operator who executes the relation: D = C _d χ F _d

The aforementioned operator makes it possible to implement the algebraic constraints representative of the malfunctions. The detector filter circuit then comprises an operator that executes the relationship: q _k = U x (γ _k - C _d x X _k ) which determines the k-th residue for the identification of the malfunction. According to the invention, it has been established that the k-th fault is decoupled from the k-th residue by a time delay equal to its detectability index. As a result, the malfunction identification circuit includes means for establishing the temporal sampling step k so that the defects and associated residues can be discernible with a given detectability index.

The two defects of the reduced model of the state representation are the defects Δ and δ _C k above and they are observable at the aforementioned conditions of the residues. They are also decoupled from each other.

In a particular embodiment, the malfunction identification circuit executes the method of the invention on the defect δ _C k-

Referring again to FIG. 1, block 7 which carries out the residue processing operation is now described. In a particular embodiment, the residue processing block 7 comprises a digital filter which comprises means for increasing the signal-to-noise ratio in order to separate the disturbances from the measurement noises. To this end, the digital residue processing filter comprises: a means for attenuating the amplitude of the unwanted signal by a certain factor; means for maintaining the slope (first order derivative) of the disturbance.

Referring to Figure 7, there is described a particular embodiment of such a digital filter having two circuits 97 and 99 realizing a looped integrator negative feedback. It comprises an integrating circuit 99 order "1" whose input u is connected to a first output of an amplifier circuit 97 whose second output can produce the filtered residue signal qf. The amplifier 97 comprises a first input which receives the residual signal q coming from the block 5 and a second input which is connected to the output iu of the integrator 99 of order "1".

In a particular embodiment, the integrator 99 comprises a circuit executing the Euler method which, at the iteration P + 1, uses the preceding sample p according to a relation of the form:

/ w [p + 1] = iu [jp] + Te ^* u [k] for a given integration time Te. In a particular embodiment, the loop gain amplifier 97 comprises an operator performing the function F () defined by:

F (x) = Kj; if ai-1 <| x | > a-, with i> 0 For x = q - iu

Preferably, the amplifier 97 comprises a means for eliminating the values Kj which do not satisfy the relation Kj <Kj for i <j.

In addition, the amplifier 97 has an output circuit which executes the linear operation: qf = G97 * iu + Offset in which G97 is a given gain and Offset a predetermined value, which may be zero.

In a particular embodiment, the amplification circuit 97 cooperates with a circuit for adapting the dynamics of the filtered signal qf according to the vehicle. The adaptation circuit then comprises means for analyzing the dynamics of the filtered signal qf and for deducing therefrom an adapted value G97 of the gain. In a particular embodiment, the output of the digital filter of the residue is connected to a comparator (not shown) at a predetermined threshold. The output of the threshold comparator produces a Boolean signal of presence or absence of detected fault. In a particular embodiment, the threshold value is produced by a threshold adapter circuit as a function of the desired sensitivity of the detectability of the malfunctions.

In a particular embodiment, the output of the digital filter filter of the residue is connected to a circuit producing a signal proportional to the intensity of the identified fault.

The malfunction identification circuit of the invention then comprises, as has been described in FIG. 1, a block 9 for detecting a malfunction. Block 9 is connected especially at the output of the filtering circuit of the residue. I l comprises a comparator (not shown) at a predetermined threshold. The output of the threshold comparator produces a Boolean signal of presence or absence of detected malfunction. In a particular embodiment, the threshold value is produced by a threshold adapter circuit as a function of the desired sensitivity of the detectability of the malfunctions.

The malfunction identification circuit of the invention finally comprises, as has been described in FIG. 1, a fault signaling block 1 1. Preferably, the block 1 1 comprises means for visualizing the presence of a fault on the electric torque of the electric steering assist motor and means for displaying the intensity or level of the fault. As is apparent from the foregoing, the identified and / or measured defect is perfectly decoupled from the disturbances taken into account in the retained state representation model.

Of course, the present invention has been described mainly for the salt fault on the electric torque of the electric motor assistance. But other defects are also detectable.

In Figure 6, there is shown a particular embodiment of a detector filter of the malfunction identification circuit and for producing a decoupled residue noted gammak and which has been described above under the notation γk and residues noted Residue delta M for a perturbation of moment type Ô _M and delta residual Ce for a disturbance on the electrical torque δce, which are respectively transmitted to residue display devices 58, 60 and 62. The detector filter has four inputs 50 to 56 which are respectively:

- Command_couple_moteur which corresponds to the instruction K above; Thetatp sensor which is the signal for measuring the speed of rotation of the electric motor or the steering of the wheels as has been explained above;

- Thetat sensor which is the measurement signal of the instantaneous angle of the engine or wheel steering; and

- This motor is the measurement signal of the current or the electric torque of the electric assistance motor.

The aforementioned signals are gathered on a discrete state-of-state state carrier (64), whose output is connected to a first input of a subtractor 68. The negative input of the subtractor 68 is connected to the output of an amplifier 66 which will be described later. The output of the subtracter 68 is connected to the input of an amplifier 70 which executes the summation of the vector input 50 uvec processed by the filter and produces the gammak decoupled residue on the aforementioned visualization 58. The output of the subtractor 68 is also connected to the inputs of two amplifiers 72 and 76. The amplifier 76 executes the product of the processed input with the gain kk and transmits its output to a gatherer 80. The gatherer 80 is also connected to an amplifier 78 Bd * u of the measurement signal from the setpoint 50 and the output of an amplifier 84 Ad * uvec whose input is common with that of the aforementioned amplifier 66 which performs the transfer Cd * uvec.

The output of the formatter 80 produces a sampling as described above Xch (k + 1) which is connected to an input of a transfer delay circuit 82 1 / z. The output Xch (k) of the circuit 82 is connected to the inputs of the above-mentioned amplifier 66 of gain Cd * uvec and of the aforementioned amplifier 84 of gain Ad * uvec.

Reference will now be made to FIGS. 5a to 5d which show the elaboration of the identification of a dysfunction from the development of a residue as described.

FIG. 5a shows the signal coming from the output of the detector filter in the identification circuit of dysfunction of the invention. This signal, once processed by the block 7 of FIG. 1, makes it possible to obtain in FIG. 5b an average curve R which has a positive slot between the instants t1 and t2 and then a negative slot between the instants t3 and t4. In Figure 5c, the Boolean fault detection signal produced after the digital filtering of the residue signal is then shown.

Finally, in FIG. 5d, there is shown the display of the amplitude of the decoupled faults of the possible disturbances and representable on the display device of the block 1 1 of FIG.

Claims

REVEN D ICATIONS

1 - Method for identifying malfunction of an electric motor for a vehicle power steering, characterized in that it comprises a preliminary step for generating a dynamic state of the vehicle model, with the power steering system, then a step to identify faults acting on the electric motor of a steering mechanism.

2 - Process according to claim 1, characterized in that the preliminary step comprises a step for establishing a state representation in which the time derivative of a state vector of the represented system is identified with the sum of the state instantaneous representation of the system shown, the electric state of the electric motor and the instantaneous state of the disturbances applied to the system shown.

3 - Process according to claim 2, characterized in that the step for establishing a state representation comprises a step for establishing at least one of the following disturbances:

a momentary perturbation on the chassis of the vehicle; a stress-type disturbance on the chassis of the vehicle; and or

a disturbance of the current type or of the electric torque type on the electric motor of the electric power steering.

4 - Process according to claim 2 or 3, characterized in that the step for establishing a state representation comprises a step for establishing the instantaneous state of the system represented on the basis of parameters for specifying the representation space of the system. state among which:

• yaw rate;

• the angle of drift; • the angular rate of longitudinal deflection;

• the longitudinal steering angle;

• the electric current flowing through the electric machine of the steering mechanism; The time derivative of the electrical current flowing through the electric machine of the steering mechanism. 5 - Process according to claim 2, characterized in that the step for identifying the faults acting on the electric steering motor comprises: a first step for detecting the instantaneous values of the state vector of the system shown as well as the torque setpoint engine;

a sixth step for exploiting the detection of at least one fault of the fifth step, in particular by signaling the detected fault as the display of an alarm. 6 - Electric vehicle power steering of the kind comprising at least one electric steering assist motor decoupled from a steering mechanism, characterized in that the electric power steering cooperates with a malfunction identification circuit integrating a representation model of states and a residue treatment means for deriving the identification of a malfunction of the electric steering assist motor. 7 - Steering according to claim 6, characterized in that, to implement the fault identification method of the electric motor assistance of claim 5, it comprises: - a block (1) for detecting sensors and shaping the signals from the sensors associated with the operation of an electric power steering;

a block (3) in which the signal coming from the sensors is processed so as to produce a vector signal for processing sensors;

a block (5) which executes an isolator filtering on the basis of the sensor processing vector signal and provides at its output a filtered signal isolated from said vector signal;

a residue treatment block (7) according to a method of functional analysis in the state space also derived from a modeling of the dynamics of the electric power steering system;

a block (9) applying a decision algorithm whose output signal is a vector decision signal which is transmitted to the input of the last block; and

an output block (1 1) in which the decision signal is processed so as to provide signaling of a malfunction or defect.

8 - Direction according to claim 7, characterized in that the detector filter circuit (7) comprises means for implementing a state representation model.

9 - Steering according to claim 8, characterized in that it comprises a reduction means of the complete model taking into account the electrical dynamics of the engine and the movement of the equivalent mass of the entire steering mechanism composed of the electric motor assistance, rack and steering wheels. 10 - Direction according to claim 8 or 9, characterized in that the detector filter (5) comprises a discrete linear filter for detecting defects of the electric motor.

1 - Direction according to claim 10, characterized in that the discrete linear filter comprises a means for calculating residues associated with the disturbances intended to identify the malfunctions on the basis of the observable states of the state representation model.

12 - Direction according to claim 1 1, characterized in that the discrete linear filter comprises means for decoupling the residues of each identifiable dysfunction.

13 - Direction according to claim 12, characterized in that the decoupling means comprises a discrete filter synthesis circuit based on a determined set of algebraic constraints representative of malfunctions.

14 - Direction according to claim 13, characterized in that the discrete synthesis circuit comprises a means for discretizing the input signals described in the reduced state representation on the basis of a sampling clock k. 15 -Direction according to any one of claims 9 to 14, characterized in that the model reduced by state representation is determined by the data of matrices with constant coefficients or time evolution (A _r , B _r , F _r , VS-).

16 - Direction according to claim 15, characterized in that the malfunction identification circuit comprises sampling circuits controlled by the clock k, the discrete filter establishing the discretized reduced model determined by the data of the matrices with constant coefficients or with time evolution sampled according to k (A _d , B _d , F _d , C _d ) used in the discretized reduced model.

17 - Direction according to claim 16, characterized in that the means for forming the residues concerned for the identification of the malfunction of the electric motor comprises an operator that executes the relation: D = C _d χ F _d and that it comprises means for implementing algebraic constraints representative of the malfunctions.

18 - Direction according to claim 7, characterized in that the block (7) for processing residues comprises a digital filter which comprises means for increasing the signal-to-noise ratio to separate the disturbances of measurement noise.

19 - Direction according to claim 18, characterized in that the digital residue treatment filter comprises:

means for attenuating the amplitude of the unwanted signal by a certain factor; means for maintaining the slope (first order derivative) of the disturbance.

20 - Direction according to claim 19, characterized in that the digital filter comprises two circuits (97, 99) providing a looped integrator with negative feedback. 21 - Steering according to at least one of claims 15 to

20, characterized in that the output of the digital filter of the residue is connected to a comparator at a predetermined threshold to produce a Boolean signal of presence or absence of detected fault. 22 - Direction according to claim 20, characterized in that the output of the digital filter filter residue is connected to a circuit producing a signal proportional to the intensity of the identified defect.

23 - Direction according to any one of claims 18 to 22, characterized in that the malfunction identification circuit of the invention comprises a malfunction detection block 9, connected in particular to the output of the residue filtering circuit which has a comparator at a threshold predetermined, so that the output of the threshold comparator produces a Boolean signal of presence or absence of detected malfunction.

24 - Steering according to claim 10, characterized in that the block (1 1) of fault signaling comprises a means of visualizing the presence of a fault on the electric torque of the electric steering assistance motor and a means of visualization of the intensity or level of the fault.