Patent Application
20230068905
The invention relates to the field of controlling and monitoring multiphase electric motors which operate with an inverter.
Many applications use electric motors with multiple phases supplied with power by a DC voltage source such as a battery, through an inverter. In motor vehicles, for example, such electric motors are supplied with power by the vehicle's batteries.
In some, particularly critical applications, the operation of these electric motors is monitored with a view first of all to identifying any malfunction, and then to identifying the source of this malfunction in order to apply corrective measures or to switch to an adapted degraded operating mode.
An electric motor controlling the steering of a motor vehicle is an example of one of these particularly critical applications. Indeed, the loss of the steering or assisted-steering function, that is to say for controlling the angular direction of the vehicle's wheels, presents a critical risk, and this criticality is greatest when, for example, the vehicle is an autonomous vehicle.
In these vehicle steering or assisted-steering applications, any malfunction in the inverter-motor assembly must be detected and identified in a very short time in order to be able to take measures appropriate for the identified malfunction before this malfunction has any effect on the driving of the vehicle.
Malfunction detection methods capable, for example in a motor vehicle, of detecting the symptoms of a malfunction in the inverter-motor assembly for a particular application are known.
There are also known palliative methods allowing the implementation of a corrective solution or a degraded operating mode allowing the function of the electric motor to continue to be carried out in safety. For example, when one of the phases of an electric power-steering motor is defective, motor control can be adapted to operate only with two of the phases of the three-phase motor, in degraded mode, thereby allowing power steering to be maintained while waiting for maintenance.
To implement such a palliative method in response to the detection of a malfunction, it is necessary to identify this malfunction, that is to say to identify precisely which electrical or mechanical member is faulty within the inverter-motor assembly. These methods for identifying a malfunction are particularly difficult to implement and it is on them that the correct implementation of palliative methods depends.
Such methods are known for identifying malfunctions in an inverter-motor assembly which use diagnostic algorithms to identify the type of failure and to discern which phase of the motor is faulty. Discerning which phase of the motor is faulty is based on an inverse digital model of the motor, which uses the voltage and current measurements taken on the motor before the occurrence of the fault. The performance of these algorithms is highly dependent on the precision of the motor parameters (resistance, inductance, motor flux). However, these parameters change according to many external factors such as temperature or current saturation. These algorithms must therefore be supplemented with advanced and sufficiently precise thermal models of the various members of the motor, with redundant temperature sensors, and with precise calibrations of the motor parameters. These known methods lead to diagnostic complexity in order to cover the entire operating range, for example, of a motor vehicle member such as power steering: high and low temperature, high speed, high torque, voltage or speed gradient, etc. These methods also require significant computing resources to run these advanced algorithms.
Other known methods for identifying malfunctions use hardware components intended to drive the power stage of the electric motor control. On detecting a hardware malfunction, such a component is capable of diagnosing which phase of the electric motor is faulty by continuously monitoring the power stage and/or in a diagnosis carried out after the electric motor has been deactivated. These methods are generally based on the injection of a weak current flowing between phases of the electric motor. Such methods require the presence of an expensive component and physically encumber the printed circuits intended for controlling the electric motor. Moreover, it is not acceptable in certain applications, such as motor vehicle power steering, to generate a current in the motor to carry out a diagnosis. Furthermore, a supervisory hardware component may be sensitive to the speed of the motor because the voltages in the phases of the motor may experience interference from the counter-electromotive force brought about by the speed of the electric motor.
The aim of an aspect of the invention is to improve the methods of the prior art for identifying a malfunction in an inverter-motor assembly.
To that end, an aspect of the invention relates to a method for identifying a malfunction in a multiphase inverter-motor assembly comprising an electric motor and an inverter having power switches distributed across branches, each phase of the electric motor being connected to a branch of the inverter by a phase switch, this method comprising a plurality of sequential diagnostic procedures each comprising the following steps:
Such a method for identifying a malfunction makes it possible to discern the origin of a failure in a very short time following the detection of this failure. The method according to an aspect of the invention is particularly suited to automotive standards and allows, for example, a malfunction to be identified within 100 ms following the occurrence of the malfunction over the entire speed range of a power-steering motor.
The method according to an aspect of the invention does not require precise knowledge of the parameters of the motor. Calibration of these parameters in production, or knowledge of their change with temperature and current, is therefore not necessary. Moreover, in operation, no temperature measurement within the electric motor or other members is necessary.
In addition, the method according to an aspect of the invention has low sensitivity to interference relating to large torque, speed or voltage gradients (the identification of the malfunction is carried out outside the current motor control phases).
The method is relatively straightforward to implement and requires little in the way of computing resources for the implementation thereof within a computer.
The cost of implementation and the space taken up on a printed circuit board are limited (only two transistors and a few resistors are required, in one embodiment).
The method is not intrusive, which ensures operating safety (no current injection is carried out on the power stage).
The method for identifying a malfunction may comprise the following additional features, alone or in combination:
Other features and advantages of aspects of the invention will become apparent from the following non-limiting description, with reference to the appended drawings, in which:
The motor-inverter assembly comprises an electric motor 1 supplied with power by an inverter 2 connected to a battery 3. In the present example, the motor 1 is a three-phase electric motor for assisting in or controlling the steering of a vehicle. The motor 1 has three phases u, v, w connecting the motor 1 to the inverter 2. A set of phase switches 7 are provided to allow each of the phases u, v, w to be interrupted.
The inverter 2 consists of power transistors distributed across three branches corresponding to the three phases u, v, w. The inverter 2 is controlled in a conventional manner for this type of application by virtue, for example, of a computer integrating a microcontroller and the other electronic elements for controlling the electric motor (the computer is not shown in the figures). This computer is suitable for implementing malfunction detection methods as well as palliative methods when a malfunction is detected and identified.
The members for controlling such an electric motor as well as the malfunction detection methods and palliative methods are known and will not be described in more detail here.
The present description relates to a method for identifying a malfunction implemented as soon as a malfunction has been detected. To implement this identification method, the inverter-motor assembly comprises a phase-biasing assembly 4 which is designed to apply a predetermined voltage to each of the phases u, v, w. This phase-biasing assembly is associated with a biasing module 5 which is designed to activate or deactivate the biasing of the phases by acting on the phase-biasing assembly 4.
The inverter-motor assembly also comprises a module 6 for configuring the inverter 2. This configuration module 6 allows the operation of the inverter 1 to be controlled according to a predetermined pulse-width modulation signal for each phase u, v, w, or, conversely, allows the inverter 1 to be deactivated.
The inverter-motor assembly further comprises:
The method for identifying a malfunction sequentially implements a plurality of diagnostic procedures all having the same general structure.
In the present example, each of these sequential diagnostic procedures comprises the following steps in particular:
In a first step E1, which is prior to the identification method, a malfunction is signaled following the implementation of a conventional malfunction detection method (not described). The malfunction identification method starts from this step E1 which may comprise, for example, the raising of a flag indicating the detection of a malfunction that needs to be identified.
In the present example, the method for identifying a malfunction comprises four sequential diagnostic procedures illustrated by steps E2, E3, E4, as well as E5a and E5b. These four sequential diagnostic procedures are executed in this order and allow, respectively:
Step E6 corresponds to the raising of a flag identifying a malfunction to allow a conventional palliative method (not described here) to take over in order to apply corrective measures or degraded operation, based on the identification of the malfunction.
Each of these sequential diagnostic procedures will now be described with reference to
In this
The inverter 2 comprises three branches corresponding to the three phases u, v, w, each of the branches comprising two power transistors, for example MOS (metal-oxide-semiconductor) transistors, connected on the one hand to ground and on the other hand to the power supply DCLink. The inverter 2 thus comprises:
Each of the branches of the inverter 2 drives one of the phases u, v, w by virtue of its high transistor 12 and its low transistor 13.
With reference to
This first diagnostic procedure starts with a first step of initially configuring the inverter which consists here in deactivating the control of the power stage, that is to say all of the transistors 12, 13 are positioned in the passive state (i.e. open) for this first diagnostic procedure.
In a second step of initially configuring the phase switches 7, these phase switches 7 are all open so as not to generate any interference in the motor 1 in the event of a failure in the power stage and so as not to experience the effects the counter-electromotive force brought about by the potential speed of the electric motor which is still rotating (since the diagnostic procedure may be carried out while the motor 1 is rotating).
In a step of biasing the phases, the biasing module 5 controls the biasing circuit 4 so as to bias all of the phases u, v, w with a predetermined diagnostic voltage Vb. The diagnostic voltage Vb is, in this example, 50% of the supply voltage DCLink of the inverter 2.
In a voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w by virtue of the voltage module 8.
In a comparison step, the voltage measurement of each phase u, v, w is compared with an expected resultant value, and In the present example, the expected resultant value is equal to the diagnostic voltage Vb. Specifically, if no transistor 12, 13 is shorted, then the voltages measured for each of the phases u, v, w must be equal to Vb (see
Conversely, if the voltage measured on one of the phases u, v, w is equal to the supply voltage DCLink, it means that a short circuit is present on the high-side transistor 12 of the phase in question. Specifically, the phase is directly connected to the voltage DCLink if such a short circuit is present. Similarly, if the voltage measured on one of the phases u, v, w is equal to zero, this means that a short circuit is present on the low-side transistor 13 of the phase in question. Specifically, the phase is directly connected to ground if such a short circuit is present.
For example, if transistor 12u is shorted, the voltages measured in the voltage measurement step for phases v and w will be equal to the diagnostic voltage Vb while the voltage measured on phase u will be equal to the supply voltage DCLink. Similarly, if transistor 13w is shorted, the voltages measured in the voltage measurement step for phases u and v will be equal to the diagnostic voltage Vb while the voltage measured on phase w will be equal to zero.
In practice, the voltage measurement step is preferably carried out with a plurality of measurements (for example 20) over a predetermined duration and by determining the average of these measurements to obtain the voltage value sought. The greater the number of these measurements, the more reliable the voltage measurement.
Similarly, the comparison step is preferably carried out by determining:
This comparison between the voltage measured on the phases u, v, w and the diagnostic voltage Vb is made for each of the phases u, v, w.
On completion of this first diagnostic procedure, if one or more transistors 12, 13 are identified as being shorted, the system raises a flag (in step E6 of
Following this first sequential diagnostic procedure, if no shorted transistor has been identified, a second sequential diagnostic procedure then takes place (step E3 of
This second diagnostic procedure has two parts executed sequentially. The first part makes it possible to diagnose an anomaly in the control of the low-side transistors 13 or a short circuit between two phases u, v, w. The second part makes it possible to diagnose an anomaly in the control of a high-side transistor 12.
For these two parts, the first diagnostic procedure starts with a first step of initially configuring the inverter which here consists in positioning the transistors 12, 13 in the passive state (i.e. open), with the exception of the transistors of one of the branches of the inverter which corresponds to a phase u, v, w referred to as the phase being tested. The three phases u, v, w will be tested in turn via three iterations of the method, each allowing one phase u, v, w to be tested. For example, in a first iteration, the first phase tested is phase u. While the transistors of the two other phases v, w are kept open, the transistors corresponding to phase u are then pulse-width modulation controlled with a low duty cycle, i.e. one much lower than 50% (for example 10%), for the first part of the diagnostic procedure, and with a high duty cycle, i.e. one much higher than 50% (for example 90%), for the second part of the diagnostic procedure.
In a second step of initially configuring the phase switches 7, these phase switches 7 are all open.
In a step of biasing the phases, the biasing module 5 controls the biasing circuit 4 so as to bias all of the phases u, v, w with a predetermined diagnostic voltage Vb. The diagnostic voltage Vb is, in this example, 50% of the supply voltage DCLink of the inverter 2.
In a voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w by virtue of the voltage module 8.
In a comparison step, the voltage measurement of each phase u, v, w is compared with an expected resultant value, and The expected resultant value for each phase is different for the first part of the diagnostic procedure and for the second part of the diagnostic procedure.
For the first part of this second diagnostic procedure:
The following faults may be identified in this first diagnostic part when the voltages measured on phases u, v, w do not correspond to the expected resultant values:
For example, if phase v is being tested:
Three iterations of this first part of the diagnostic procedure allow each phase to be switched to the role of phase being tested, and thus allow the above malfunctions to be identified for all of the phases u, v, w.
For the second part of this second diagnostic procedure:
The following fault may be identified in this second diagnostic part when the voltages measured on phases u, v, w do not correspond to the expected resultant values:
For example, if phase v is being tested:
Three iterations of this second part of the diagnostic procedure allow each phase to be switched to the role of phase being tested, and thus allow high-side transistor control malfunctions to be identified for all of the phases u, v, w.
On completion of this second diagnostic procedure, if the control of one or more transistors 12, 13 is identified as faulty, or if a short circuit between two phases is identified, the system raises a flag (in step E6 of
Following this second sequential diagnostic procedure, if no phase short circuit or faulty transistor control has been identified, a third sequential diagnostic procedure then takes place (step E4 of
This third diagnostic procedure starts with a first step of initially configuring the inverter which consists here in deactivating the control of the power stage, that is to say all of the transistors 12, 13 are positioned in the passive state (i.e. open) for this third diagnostic procedure.
In a second step of initially configuring the phase switches 7, these phase switches 7 are all positioned and kept in the closed state in order to be able to diagnose the phases u, v, w between the phase switches and the motor.
In a step of biasing the phases, the biasing module 5 controls the biasing circuit 4 so as to bias all of the phases u, v, w with a predetermined diagnostic voltage Vb. The diagnostic voltage Vb is, in this example, 50% of the supply voltage DCLink of the inverter 2.
In a voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w by virtue of the voltage module 8. In this step, an average of the three phase voltages is also calculated. The three measured voltages are added and the result is divided by three, which makes it possible to obtain a voltage value resulting from the counter-electromotive forces, and therefore independent of the variation in the counter-electromotive forces brought about by the rotation of the motor.
In a comparison step, the average of the three phase voltages is compared with an expected resultant value. In the present example, the expected resultant value is equal to the diagnostic voltage Vb.
Conversely, if the average of the three phase voltages is lower than the voltage Vb (potentially minus a safety threshold), or if the average of the three phase voltages is lower than 0 V plus a predetermined threshold, this means that a short circuit is present between at least one phase and ground. Similarly, if the average of the three phase voltages is higher than the voltage Vb (potentially plus a safety threshold), or if the average of the three phase voltages is higher than DCLink minus a predetermined threshold, this means that a short circuit is present between at least one phase and the supply voltage DCLink.
On completion of this third diagnostic procedure, if one or more phases, on the motor side, are identified as being shorted to ground or the supply voltage, the system raises a flag (in step E6 of
Following this third sequential diagnostic procedure, if no short circuit has been identified on the motor phases, a fourth sequential diagnostic procedure then takes place (step E5a and E5b of
There are two variants of this fourth diagnostic procedure depending on the electrical rotational speed. The electrical rotational speed corresponds to the rotational speed of the motor multiplied by a factor relating to the number of pairs of motor poles. For example, for a motor with four pole pairs, the ratio of the electrical rotational speed to the motor rotational speed is 4. The first variant is implemented in the case of an electrical rotational speed lower than 210 rad/s for example, and the second variant is implemented in the case of a high electrical rotational speed, for example higher than 210 rad/s. As a variant, the speed threshold for selecting the first variant or the second variant may be dynamically adjusted as a function, in particular, of the available supply voltage.
In its first variant (slow motor speed), the fourth diagnostic procedure starts with a first step of initially configuring the inverter which here consists in positioning the transistors 12, 13 in the passive state (i.e. open), with the exception of the transistors of one of the branches of the inverter which corresponds to a phase u, v, w referred to as the phase being tested. The three phases u, v, w will be tested in turn via three iterations of the method, each allowing one phase u, v, w to be tested. For example, in a first iteration, the first phase tested is phase u. While the transistors of the two other phases v, w are kept open, the transistors corresponding to phase u are then pulse-width modulation controlled in a way that is chosen to be preferably away from 50%, i.e. with either a high value (for example 80%) or a low value (for example 20%).
In a second step of initially configuring the phase switches 7, these phase switches 7 are all positioned and kept in the closed state.
In a step of biasing the phases, the biasing module 5 controls the biasing circuit 4 so as to bias all of the phases u, v, w with a predetermined diagnostic voltage Vb. The diagnostic voltage Vb is, in this example, 50% of the supply voltage DCLink of the inverter 2.
In a voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w by virtue of the voltage module 8.
In a comparison step, the voltage measurement of each phase u, v, w is compared with an expected resultant value. The expected resultant value for the phase being tested is equal to the supply voltage DCLink multiplied by the duty cycle. The expected resultant value for the two other phases is equal to the preceding voltage (supply voltage DCLink multiplied by the duty cycle) plus the counter-electromotive force of the phase in question.
Conversely, if the phase being tested is interrupted at the level of the motor, the voltage measured on the phase being tested is equal to the supply voltage DCLink multiplied by the duty cycle while the voltage measured on the two other phases is equal to the diagnostic voltage Vb plus the counter-electromotive force of the phase in question.
For example, if phase u is being tested:
The counter-electromotive force of each phase may be obtained, for example, through calculation. Specifically, the amplitude of the counter-electromotive force voltage brought about by the rotational speed of the motor is equal to the speed of the motor multiplied by the flux of the motor multiplied by the square root of 3. The speed of the motor is measured by the speed module. Alternatively, the counter-electromotive force of each phase may be estimated with a threshold value since this first variant is applied at low rotational speeds.
Three iterations of this first variant of the diagnostic procedure (slow motor speed) allow each phase to be switched to the role of phase being tested, and thus allow the above malfunction to identified for all of the phases u, v, w.
On completion of this first variant of the fourth diagnostic procedure, if one or more phases, on the motor side, are identified as being interrupted, the system raises a flag (in step E6 of
In its second variant (fast motor speed), the fourth diagnostic procedure starts with a first step of initially configuring the inverter which consists here in deactivating the control of the power stage, that is to say all of the transistors 12, 13 are positioned in the passive state (i.e. open) for this second variant of the fourth diagnostic procedure.
In a second step of initially configuring the phase switches 7, these phase switches 7 are all positioned and kept in the closed state.
In a step of biasing the phases, the biasing module 5 controls the biasing circuit 4 so as to bias all of the phases u, v, w with a predetermined diagnostic voltage Vb. The diagnostic voltage Vb is, in this example, 50% of the supply voltage DCLink of the inverter 2.
In a voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w by virtue of the voltage module 8. In this example, a sampling of N voltage values (for example 20) takes place simultaneously on each phase. A sample “Phase u(n)” thus designates a voltage measurement taken on phase u at time n, the sample “Phase v(n)” designates the voltage measurement taken on phase v at the same time n, and the sample “Phase w(n)” designates the voltage measurement taken on phase w at the same time n. If, for example, a sampling of two voltage measurements is taken on the phases u, v, w (N=2), two groups of three voltage measurements will be available:
In a step of measuring the rotational speed of the motor, a sampling of N speed values takes place simultaneously with the voltage measurements of the preceding step. Each speed measurement sample, denoted MotSpdEl(n), therefore corresponds to a motor speed measurement taken simultaneously with the Phase u(n), Phase v(n), and Phase w(n) voltage samples.
In a comparison step, the voltage measurement of each phase u, v, w is compared with an expected resultant value. This comparison step implements a comparison between a representative voltage Vdiag determined for each phase u, v, w and a threshold voltage Vseuil determined as indicated below. The principle is that the expected resultant values are:
Thus, if according to the example above, phase v is being tested and if this phase is interrupted at the level of the motor, then the voltage value measured on this phase v will be substantially equal to the diagnostic voltage Vb and the sum of the counter-electromotive force of phase u and of the counter-electromotive force of phase w will be zero.
In practice, the representative voltage Vdiag for a phase u, v, w is preferably calculated for each measured voltage sample n. The representative voltage Vdiag for one of the phases (called the phase being tested) and for a voltage measurement sample “n” is equal to the absolute value of the average of the voltages n of all of the phases u, v, w minus the average of the voltages n of the two phases which are not being tested. The calculation of this representative voltage Vdiag is written as:
This value VdiagPhX(n) is therefore calculated for each phase u, v, w in three iterations where the phases u, v, w are sequentially designated as the phase being tested.
In addition, a threshold voltage is calculated for each motor speed measurement sample “n”, according to the formula:
Vseuil(n)=VseuilMin+SpdRatio×max(0,|(MotSpdEl(n))|−MotSpdElMin) [Math 2]
In the preceding formula:
The comparison step is carried out by comparing, for each phase u, v, w, the value VdiagPhX(n) and the corresponding value Vseuil(n). When VdiagPhX(n) is lower than Vseuil(n), an error is qualified for sample n, and for the phase u, v, w in question. When, for a given phase u, v, w, more than 80% of the samples n lead to a qualified error, the failure is identified, i.e. an interruption in the continuity of the motor phase of the phase in question. This level of 80% corresponds to an identification threshold which may be adjusted.
On completion of this second variant of the fourth diagnostic procedure, if a phase u, v, w is identified as interrupted at the level of the motor, the system raises a flag (in step E6 of
This voltage divider assembly comprises a logic signal input 14 connected to the microcontroller which activates or deactivates the biasing of the phases. The logic signal input 14 drives a MOS transistor 15 placed between the supply voltage DCLink and ground, via a bipolar transistor 17. When the MOS transistor 15 is controlled (with a view to biasing the phases in the phase-biasing steps), the motor phases are connected to the supply voltage DCLink by pull-up resistors 18. When the MOS transistor 15 is not controlled, the phases u, v, w are connected to ground via the pull-down resistor 19.
High resistance values are used for the resistors 18, 19 so that the assembly does not interfere with the driving of the motor phases by the power stage when the latter is active.
Any other electronic assembly allowing the diagnostic voltage Vb to be applied to the phases u, v, w may alternatively be used as a biasing module 5.
Other variant embodiments of the method for identifying a malfunction may be implemented without departing from the scope of the invention. For example, only some of the sequential diagnostic procedures described may be carried out to constitute a method for identifying a malfunction.
| Number | Date | Country | Kind |
|---|---|---|---|
| FR2001078 | Feb 2020 | FR | national |
This application is the U.S. National Phase Application of PCT International Application No. PCT/EP2020/087208, filed Dec. 18, 2020, which claims priority to French Patent Application No. 2001078, filed Feb. 4, 2020, the contents of such applications being incorporated by reference herein.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2020/087208 | 12/18/2020 | WO |
