Method for identifying a malfunction of an inverter-motor assembly

Abstract

Method for identifying a malfunction of an inverter-multiphase electric machine assembly comprising a plurality of sequential diagnostic procedures, each sequential diagnostic procedure comprising the following steps: - an initial configuration step of the inverter (2); - an initial configuration step of the phase switches (7); - a phase (u, v, w) biasing step; - a voltage measurement step in which the voltage of each phase (u, v, w) is measured; - a comparison step in which the voltage measurement of each phase (u, v, w) is compared with an expected result value; - a malfunction identification step when the voltage measurement of one phase is different from the expected result value.

Description

Technical Field

[0001] This invention relates to the field of control and monitoring of multiphase motors operating in conjunction with inverters.

[0002] Many applications use motors with multiple phases, which are powered by a DC voltage source such as a battery via an inverter. For example, in motor vehicles, such motors are powered by the vehicle's battery.

[0003] In some particularly critical applications, monitoring the operation of these motors is primarily to identify any dysfunctions, and then to identify the source of those dysfunctions in order to take corrective action or switch to an appropriate degraded operating mode.

[0004] The motor that controls the steering system of a motor vehicle is an example of these particularly critical applications. In fact, the loss of steering or assisted steering function (i.e., controlling the angle and direction of the vehicle's wheels) poses serious risks, and these dangers are most severe when it comes to vehicles such as autonomous vehicles.

[0005] In these vehicle steering or assisted steering applications, any malfunctions in the inverter-motor assembly must be detected and identified within a very short time in order to take appropriate measures to address the malfunction before it has any impact on vehicle driving. Background Technology

[0006] For example, in motor vehicles, there are known methods for detecting malfunctions in inverter-motor assemblies used for specific applications.

[0007] Mitigation methods are also known that enable the implementation of corrective solutions or degraded operating modes, thereby allowing the motor to continue to function safely. For example, when one phase of the auxiliary steering motor is defective, the motor control can be adapted to operate in degraded mode using only two phases of the three-phase motor, thereby maintaining auxiliary steering while awaiting repair.

[0008] To implement such a mitigation method that responds to the detection of a malfunction, it is necessary to identify the malfunction, that is, to accurately identify which electrical or mechanical component within the inverter-motor assembly is malfunctioning. These malfunction identification methods are particularly difficult to implement, and the correct implementation of mitigation methods is based on these methods.

[0009] Known methods for identifying faults in inverter-motor components utilize diagnostic algorithms that enable the identification of fault types and pinpointing which phase of the motor is faulty. This identification is based on an inverse digital model of the motor, using voltage and current measurements taken before the fault occurred. The performance of these algorithms largely depends on the accuracy of motor parameters (resistance, inductance, and flux flux). However, these parameters can evolve with numerous external factors such as temperature or current saturation. Therefore, these algorithms must be supplemented with advanced and sufficiently accurate thermal models of the motor components, redundant temperature sensors, and accurate calibration of the motor parameters. These known methods lead to diagnostic complexity in covering the full operating range of components in motor vehicles, such as auxiliary steering: high and low temperatures, high speeds, high torque, voltage or speed gradients, etc. These methods also require significant computational resources to operate these advanced algorithms.

[0010] Other known malfunction identification methods utilize hardware components designed to manipulate the control power stage of the motor. Upon detecting a hardware malfunction, such components can diagnose which phase of the motor is faulty by means of continuous monitoring of the power stage and / or diagnostics performed after the motor has been deactivated. These methods are typically based on the injection of a weak current flowing between the phases of the motor. Such methods require expensive components and physically result in a large printed circuit board for controlling the motor. Furthermore, in some applications, such as auxiliary steering in motor vehicles, inducing current in the motor to perform diagnostics is unacceptable. Additionally, the monitoring hardware components may be sensitive to motor speed, as the voltage in each phase of the motor can be disturbed by the back electromotive force caused by the motor speed. Summary of the Invention

[0011] The purpose of this invention is to improve the existing method for identifying functional faults in inverter-motor components.

[0012] Therefore, the present invention relates to a method for identifying functional faults in an inverter-multiphase motor assembly, the inverter-multiphase motor assembly including a motor and an inverter, the inverter having power switches distributed on branches, each phase of the motor being connected to a branch of the inverter via a phase switch, the method comprising multiple sequential diagnostic processes, each sequential diagnostic process including the following steps:

[0013] - Initial configuration steps for the inverter, wherein control selection is performed for each branch of the inverter, the selection being made from a group including: fixing the power switch to the off mode; pulse width modulation control based on a predetermined diagnostic duty cycle of the power switch;

[0014] - Initial configuration steps for phase switches, wherein a state selection is performed for each phase switch, which is between a closed state and an open state;

[0015] - Phase biasing step, in which a predetermined diagnostic voltage is applied to each branch of the inverter;

[0016] - Voltage measurement steps, in which the voltage of each phase is measured;

[0017] - Comparison step, in which the voltage measurement value of each phase is compared with the expected result value;

[0018] - Fault identification step: A fault is identified when the voltage measurement of a phase differs from the expected result.

[0019] This method of malfunction identification enables the source of the malfunction to be identified within a very short time after it is detected. The method according to the invention is particularly suitable for motor vehicle standards and enables the identification of a malfunction, for example, within 100 ms after its occurrence, across the entire speed range of the auxiliary steering motor.

[0020] The method according to the invention does not require precise knowledge of the motor parameters. Therefore, there is no need to calibrate these parameters during production, nor is it necessary to know how they evolve with temperature and current. Furthermore, no temperature measurements are required within the motor or other components during operation.

[0021] Furthermore, the method according to the invention is low in sensitivity to disturbances related to large torque, speed, or voltage gradients (in fact, the identification of functional failures is performed outside the current phase of motor control).

[0022] This method is relatively simple to implement and requires very few computing resources to implement in a computer.

[0023] The implementation cost and the space occupied on the printed circuit are limited (in one embodiment, only two transistors and a few resistors are needed).

[0024] This method is non-invasive and ensures operational safety (no current injection is performed on the power stage).

[0025] Disability identification methods may include the following additional features, either individually or in combination:

[0026] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0027] - The voltage measurement step is performed by determining the average of multiple voltage measurement samples acquired for each phase;

[0028] - This method includes a first-order diagnostic process, wherein:

[0029] - During the initial configuration steps of the inverter, all power switches are controlled to the off mode;

[0030] - In the initial configuration steps of the phase switches, all phase switches are placed in the open state;

[0031] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0032] - In the comparison step, the expected result value for each phase is approximately equal to the predetermined diagnostic voltage;

[0033] - In the fault identification step, when the voltage measurement of a phase is approximately equal to the supply voltage of the inverter, a short circuit is identified on the high-side power switch of the corresponding phase.

[0034] - In the functional failure identification step, when the voltage measurement of a phase is approximately zero, a short circuit is identified on the low-side power switch of the corresponding phase.

[0035] - This method includes a second sequential diagnostic process, wherein:

[0036] - In the initial configuration step of the inverter, all power switches except the power switch corresponding to the phase being tested are controlled to the off mode, and the power switch corresponding to the phase being tested is pulse width modulation controlled according to a predetermined diagnostic duty cycle.

[0037] - In the initial configuration steps of the phase switches, all phase switches are placed in the open state;

[0038] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0039] - In the comparison step, the expected result value of the phase being tested is approximately equal to the inverter's supply voltage multiplied by the predetermined diagnostic duty cycle, and the expected result value of the other phases is approximately equal to the predetermined diagnostic voltage.

[0040] - According to the first part of the second-order diagnostic process:

[0041] - During the initial configuration steps of the inverter, the pre-defined diagnostic duty cycle is well below 50%;

[0042] - In the malfunction identification step, a short circuit between the two phases is identified when the voltage measurement of the phase being tested is approximately equal to the voltage of the other phase that is not being tested.

[0043] - When the voltage measurement of the phase being tested is higher than the predetermined diagnostic voltage, a functional failure is identified in the control of the low-side power switch of the phase being tested.

[0044] - According to the second part of the second-order diagnostic process:

[0045] - During the initial configuration steps of the inverter, the pre-defined diagnostic duty cycle is much higher than 50%;

[0046] - In the malfunction identification step, when the voltage measurement value of the phase being tested is lower than the predetermined diagnostic voltage, a malfunction is identified in the control of the high-side power switch of the phase being tested.

[0047] - This method includes a third-order diagnostic process, wherein:

[0048] - During the initial configuration steps of the inverter, all power switches are controlled to the off mode;

[0049] - In the initial configuration steps of the phase switches, all phase switches are placed in the closed state;

[0050] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0051] - In the voltage measurement step, determine the average voltage of all phases;

[0052] - In the comparison step, the expected result of the average voltage of all phases is approximately equal to the predetermined diagnostic voltage;

[0053] - In the malfunction identification step, when the average voltage of all phases is lower than the predetermined diagnostic voltage, a short circuit is identified between at least one phase and ground, which is located between the phase switch and the motor.

[0054] - In the malfunction identification step, when the average voltage of all phases is higher than the predetermined diagnostic voltage, a short circuit is identified between at least one phase and the inverter's supply voltage, and the short circuit is located between the phase switch and the motor.

[0055] This method includes a first variation of the fourth sequential diagnostic procedure, wherein:

[0056] - In the initial configuration step of the inverter, all power switches except the power switch corresponding to the phase being tested are controlled to the off mode, and the power switch corresponding to the phase being tested is pulse width modulation controlled according to a predetermined diagnostic duty cycle.

[0057] - In the initial configuration steps of the phase switches, all phase switches are placed in the closed state;

[0058] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0059] - In the comparison step, the expected result value of the phase being tested is approximately equal to the inverter's supply voltage multiplied by a predetermined diagnostic duty cycle, and the expected result value of the other phases is approximately equal to the voltage of the phase being tested plus the voltage obtained by the motor applying back electromotive force to the phase in question.

[0060] - In the fault identification step, when the voltage of the phase being tested is approximately equal to the inverter's supply voltage multiplied by a predetermined diagnostic duty cycle, and when the voltage of other phases is approximately equal to the predetermined diagnostic voltage plus the back electromotive force of the phase involved, an open circuit is identified on the phase being tested, between the phase switch and the motor.

[0061] - This method includes a second variation of the fourth sequential diagnostic procedure, wherein:

[0062] - During the initial configuration steps of the inverter, all power switches are controlled to the off mode;

[0063] - In the initial configuration steps of the phase switches, all phase switches are placed in the closed state;

[0064] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0065] - In the comparison step, the expected result value for each phase is approximately equal to the predetermined diagnostic voltage plus the back electromotive force of the phase involved, wherein the sum of the back electromotive forces of all phases is equal to zero.

[0066] - In the functional failure identification step, when the voltage of the first phase is approximately equal to the predetermined diagnostic voltage and the sum of the back electromotive forces of the other phases is equal to zero, an open circuit is identified on the first phase, between the phase switch and the motor.

[0067] This method includes a second variation of the fourth sequential diagnostic procedure, wherein:

[0068] - During the initial configuration steps of the inverter, all power switches are controlled to the off mode;

[0069] - In the initial configuration steps of the phase switches, all phase switches are placed in the closed state;

[0070] - In the phase biasing step, the predetermined diagnostic voltage is approximately 50% of the inverter's supply voltage;

[0071] - In the measurement step, voltage measurement is performed based on a sample including a predetermined number of measurements;

[0072] - In the comparison step, the expected result value for each phase is approximately equal to the predetermined diagnostic voltage plus the back electromotive force of the phase involved, wherein the sum of the back electromotive forces of all phases is equal to zero.

[0073] - In the functional failure identification step, when the voltage of the first phase is approximately equal to the predetermined diagnostic voltage and the sum of the back electromotive forces of the other phases is equal to zero, an error is identified for the first phase;

[0074] - When the number of identified errors exceeds the identification threshold, an open circuit is identified on the first phase, between the phase switch and the motor;

[0075] - A first variant of the fourth sequential diagnostic process is performed at an electric speed of less than approximately 200 rad / s, and a second variant of the fourth sequential diagnostic process is performed at an electric speed of more than approximately 200 rad / s. Attached Figure Description

[0076] Other features and advantages of the invention will become apparent from the following non-limiting description and with reference to the accompanying drawings, wherein:

[0077] Figure 1 An inverter-motor assembly adapted to implement the method according to the invention is shown schematically;

[0078] Figure 2 This is a diagram illustrating the method according to the present invention;

[0079] Figure 3 Showing more details Figure 1 Inverter-motor assembly;

[0080] Figure 4 It shows the use of Figure 1 and Figure 3 The phase bias circuit (montage) of the inverter-motor assembly. Detailed Implementation

[0081] Figure 1 The inverter-motor assembly and the elements enabling the implementation of the functional failure identification method according to the invention are schematically shown.

[0082] The inverter-motor assembly includes a motor 1, which is powered by an inverter 2 connected to a battery 3. In this example, motor 1 is a three-phase motor that enables vehicle steering or vehicle steering control. Motor 1 has three phases, u, v, and w, connecting motor 1 to inverter 2. A set of phase switches 7 is provided to enable the disconnection of each of phases u, v, and w.

[0083] Inverter 2 consists of power transistors distributed on three branches corresponding to the three phases u, v, and w. Inverter 2 is operated in a conventional manner for this type of application by means of a computer (not shown in the figure) integrating, for example, a microcontroller and other electronic components adapted to control the motor. The computer is adapted to implement malfunction detection methods and mitigation methods when malfunctions are detected and identified.

[0084] The components used to control such motors, as well as methods for detecting and mitigating malfunctions, are known and will not be described in further detail here.

[0085] This description relates to a malfunction identification method implemented after a malfunction is detected. To implement such an identification method, the inverter-motor assembly includes a phase bias circuit 4 adapted to apply a predetermined voltage to each of phases u, v, and w. This phase bias circuit is associated with a bias module 5 adapted to activate or deactivate phase bias by acting on the phase bias circuit 4.

[0086] The inverter-motor assembly also includes a configuration module 6 for configuring the inverter 2. This configuration module 6 enables 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, enables the inverter 1 to be disabled.

[0087] The inverter-motor assembly also includes:

[0088] - Voltage module 8, which enables the measurement and acquisition of the voltage of each phase u, v, w;

[0089] - Speed ​​module 10, which is adapted to measure the speed of the motor;

[0090] - Identification module 9, which enables the identification of which component of the inverter-motor assembly is causing the malfunction based on elements provided by voltage module 8 and speed module 10.

[0091] Figure 2 The diagram schematically illustrates a series of sequential diagnostic processes that constitute a method for identifying functional impairments.

[0092] The functional impairment identification method sequentially implements multiple diagnostic procedures, all of which have the same overall structure.

[0093] In this example, each of these sequential diagnostic procedures specifically includes the following steps:

[0094] - The initial configuration of the inverter is performed by configuration module 6;

[0095] - Initial configuration of phase switch 7;

[0096] - The biasing module 5 is responsible for activating the biases of phases u, v, and w according to the predetermined voltage.

[0097] - The steps of measuring and acquiring the voltage of each phase are performed by voltage module 8;

[0098] - Comparison step, in which the measured voltage values ​​of phases u, v, w are compared with the expected result values;

[0099] - Fault identification step, namely, identifying which component of the inverter-motor assembly is the source of the fault through identification module 9.

[0100] Figure 2 An identification method is shown as a series of various sequential diagnostic processes, each of which has the above-described overall structure. This series of various sequential diagnostic processes is performed according to the sequence of components that enable control of the inverter-motor assembly until one or more components are identified as the cause of the malfunction.

[0101] In the first step E1 prior to the identification method, a signal is sent indicating a functional disorder after the implementation of a conventional functional disorder detection method (not described). The functional disorder identification method is initiated based on this step E1, which may include, for example, activating a flag indicating that a functional disorder requiring identification has been detected.

[0102] In this example, the functional impairment identification method includes four sequential diagnostic procedures illustrated by steps E2, E3, E4, and E5a and E5b. These four sequential diagnostic procedures are executed in this order and respectively enable:

[0103] - Step E2: Identify a short circuit in one of the inverter's power transistors;

[0104] - Step E3: Identify a short circuit between two phases u, v, w, or identify a power transistor that is normally open, or identify a transistor control malfunction;

[0105] - Step E4: Identify short circuits between phases u, v, w of the motor and ground, or between phases u, v, w and the power supply DCLink.

[0106] - Steps E5a and E5b: Identify the electrical continuity interruption that causes a permanent disconnection of one of the phases u, v, w of the motor.

[0107] Step E6 corresponds to activating the flag of the identified dysfunction based on the identification of the dysfunction, so that conventional mitigation methods (not described here) can take over to take corrective measures or downgrade the operation.

[0108] Now refer to Figure 3Describe each of these sequential diagnostic processes, Figure 3 The main components of inverter-motor assembly 1 are shown.

[0109] In Figure 3 In the three-phase motor 1, there are three windings 11. The phase switch 7 consists of three switches 7u, 7v, and 7w, each of which is adapted to disconnect one of phases u, v, and w.

[0110] Inverter 2 includes three branches corresponding to the three phases u, v, and w. Each branch includes two power transistors, such as MOS (Metal Oxide Semiconductor) transistors. These two power transistors are connected to ground on one side and to the power supply DC-Link on the other. Inverter 2 therefore includes:

[0111] - Three high-side transistors, 12u, 12v, and 12w, are connected to the supply voltage (here, the battery voltage DCLink); and

[0112] - Three low-side transistors, 13u, 13v, and 13w, are grounded.

[0113] Each branch of inverter 2 controls one of phases u, v, w by means of its high-side transistor 12 and low-side transistor 13.

[0114] refer to Figure 2 When implementing the malfunction identification method, the first sequential diagnostic process E2 is executed first to identify whether a short circuit exists on one of transistors 12 and 13.

[0115] The first diagnostic process begins with the first step of the initial inverter configuration, where the control of the power stage is disabled, i.e., during the first diagnostic process, all transistors 12 and 13 are placed in a passive state (and thus disconnected).

[0116] In the second step of the initial configuration of phase switches 7, all of these phase switches 7 are disconnected so as not to generate any interference in motor 1 in the event of a power stage failure, and so as not to be affected by the back electromotive force caused by the possible speed of the still rotating motor (in fact, the diagnostic process can be performed while motor 1 is rotating).

[0117] In the phase biasing step, biasing module 5 controls biasing circuit 4 to bias all phases u, v, and w with a predetermined diagnostic voltage Vb. In this example, the diagnostic voltage Vb is 50% of the supply voltage DCLink of inverter 2.

[0118] In the voltage measurement step, the microcontroller uses voltage module 8 to acquire the voltage of each phase u, v, w.

[0119] In the comparison step, the voltage measurements for each phase u, v, w are compared to the expected result value. In this example, the expected result value is equal to the diagnostic voltage Vb. In reality, if none of transistors 12 and 13 were short-circuited, the measured voltages for each phase u, v, w should be equal to Vb (see...). Figure 3 This is because there is no connection between phases u, v, w and ground or the DCLink power supply.

[0120] Conversely, if the voltage measured on one of phases u, v, or w is equal to the supply voltage DCLink, this indicates a short circuit in the high-side transistor 12 of the affected phase. In effect, if such a short circuit exists, the phase is directly connected to the DCLink voltage. Similarly, if the voltage measured on one of phases u, v, or w is zero, this indicates a short circuit in the low-side transistor 13 of the affected phase. In effect, if such a short circuit exists, the phase is directly grounded.

[0121] For example, if transistor 12u is short-circuited, the voltage measured for phases v and w in the voltage measurement step will be equal to the diagnostic voltage Vb, while the voltage measured for phase u will be equal to the supply voltage DCLink. Similarly, if transistor 13w is short-circuited, the voltage measured for phases u and v in the voltage measurement step will be equal to the diagnostic voltage Vb, while the voltage measured for phase w will be zero.

[0122] In practice, the voltage measurement procedure is preferably performed as follows: multiple measurements (e.g., 20 times) are taken over a predetermined time period, and the average of these measurements is determined to obtain the desired voltage value. The more measurements taken, the more reliable the voltage measurement will be.

[0123] Similarly, the comparison step is preferably performed by determining the following:

[0124] - If the voltage measured on a phase is centered on the diagnostic voltage Vb (here, 50% of the DCLink value) and includes a margin of error relative to Vb, then the voltage is considered equal to Vb (no short circuit).

[0125] - If the voltage measured on a phase is significantly higher than the diagnostic voltage Vb (here, 50% of the value of DCLink), or significantly higher than the range of voltage Vb defined above, then the measured voltage is considered equal to DCLink (short-circuited on high-side transistor 12); or

[0126] - If the voltage measured on a phase is much lower than the diagnostic voltage Vb, or much lower than the range of voltage Vb defined above, then the measured voltage is considered to be equal to zero (short circuit on low-side transistor 13).

[0127] For each of phases u, v, and w, perform a comparison between the voltage measured on phase u, v, and w and the diagnostic voltage Vb.

[0128] At the end of this first diagnostic procedure, if one or more transistors 12, 13 are identified as short-circuited, the system activates an indicator that the malfunction is a short circuit and identifies the transistors involved (in...). Figure 2 Step E6).

[0129] If no transistor short circuits are detected after this first sequential diagnostic process, the second sequential diagnostic process is then performed. Figure 2 Step E3). This second diagnostic process enables the identification of the following functional defects: short circuits between the two phases u, v, w; functional defects in the control electronics of transistors 12 and 13; and transistors stuck in the open state.

[0130] The second diagnostic process is broken down into two sequentially executed parts. The first part enables the diagnosis of control anomalies in the low-side transistor 13 or short circuits between the two phases u, v, and w. The second part enables the diagnosis of control anomalies in the high-side transistor 12.

[0131] For both parts, the first diagnostic process begins with the initial step of configuring the inverter, where transistors 12 and 13, except for those in one branch of the inverter corresponding to a phase u, v, w, referred to as the phase under test, are placed in a passive state (i.e., disconnected). The three phases u, v, w are tested sequentially using three iterations of this method, each iteration allowing one phase u, v, w to be tested. For example, in the first iteration, the first phase under test is phase u. While the transistors in the other two phases v and w remain disconnected, the transistor corresponding to phase u is then controlled according to pulse width modulation (PWM) control, which has a low duty cycle (e.g., much lower than 50%, e.g., 10%) in the first part of the diagnostic process and a high duty cycle (e.g., much higher than 50%, e.g., 90%) for the second part of the diagnostic process.

[0132] In the second step of the initial configuration of phase switches 7, all of these phase switches 7 are disconnected.

[0133] In the phase biasing step, biasing module 5 controls biasing circuit 4 to bias all phases u, v, and w with a predetermined diagnostic voltage Vb. In this example, the diagnostic voltage Vb is 50% of the DCLink supply voltage of inverter 2.

[0134] In the voltage measurement step, the microcontroller uses voltage module 8 to acquire the voltage of each phase u, v, w.

[0135] In the comparison step, the voltage measurements for each phase (u, v, w) are compared with the expected results. The expected results for each phase differ between the first and second parts of the diagnostic process.

[0136] For the first part of this second diagnostic procedure:

[0137] - The expected result value of the phase being tested is equal to the supply voltage DCLink multiplied by the low duty cycle;

[0138] - The expected values ​​for the other two phases are equal to the diagnostic voltage Vb.

[0139] When the measured voltages on phases u, v, and w are inconsistent with the expected values, the following faults can be identified in this first part of the diagnostic process:

[0140] - If the voltage measured on the phase being tested is higher than the supply voltage DCLink multiplied by the low duty cycle (possibly plus a safety threshold), a functional failure of the low-side transistor control corresponding to the phase being tested is identified.

[0141] - If the voltage measured on one of the other two phases is equal to the voltage measured on the phase being tested, which is equal to the supply voltage DCLink multiplied by the low duty cycle, then a short circuit is identified between the phase being tested and one of the other two phases.

[0142] For example, if the test is performed on phase v:

[0143] - The value on phase u is voltage Vb, the value on phase v is DCLink multiplied by the low duty cycle, and the value on phase w is Vb, which corresponds to the expected result value, and no fault was detected;

[0144] - The value on phase u is the DCLink voltage multiplied by the low duty cycle, the value on phase v is the DCLink voltage multiplied by the low duty cycle, and the value on phase w is Vb, which corresponds to the identification of a short circuit between phase u and phase v;

[0145] - A value of Vb on phase u, a value of DCLink multiplied by a low duty cycle on phase v, and a value of Vb on phase w correspond to identifying a functional failure in the 13V control of the transistor.

[0146] The three iterations of this first part of the diagnostic process enable each phase to switch sequentially to the role of the phase being tested, thereby enabling the identification of the aforementioned functional impairments in all phases u, v, and w.

[0147] For the second part of this second diagnostic procedure:

[0148] - The expected result value of the phase being tested is equal to the supply voltage DCLink multiplied by the high duty cycle;

[0149] - The expected values ​​for the other two phases are equal to the diagnostic voltage Vb.

[0150] When the measured voltages on phases u, v, and w are inconsistent with the expected values, the following faults can be identified in this second part of the diagnostic process:

[0151] - If the voltage measured on the phase being tested is lower than the supply voltage DCLink multiplied by the high duty cycle (possibly minus a safety threshold), a functional failure of the high-side transistor control corresponding to the phase being tested is identified.

[0152] For example, if the test is performed on phase v:

[0153] - The value on phase u is voltage Vb, the value on phase v is DCLink multiplied by the high duty cycle, and the value on phase w is Vb, which corresponds to the expected result value, and no fault was detected;

[0154] - The value on phase u is the voltage Vb, the value on phase v is the supply voltage DCLink multiplied by the high duty cycle, and the value on phase w is Vb, which corresponds to the identification of a functional failure in the 12V control of the transistor.

[0155] The three iterations of this second part of the diagnostic process enable each phase to switch sequentially to the role of the phase being tested, thereby enabling the identification of control malfunctions of all phase u, v, w high-side transistors.

[0156] At the end of this second diagnostic process, 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 activates a flag indicating what the malfunction is and which transistor may be involved (in...). Figure 2 Step E6).

[0157] If no phase short circuit or transistor control fault is identified after this second sequential diagnostic process, the third sequential diagnostic process is then performed. Figure 2 Step E4). This third diagnostic procedure aims to identify:

[0158] - Any short circuit between phase and ground (located between the phase switch and the motor); and

[0159] - Any short circuit between the phase and the power supply (the short circuit is located between the phase switch and the motor).

[0160] The third diagnostic process begins with the first step of the initial inverter configuration, where the control of the power stage is disabled, i.e., during the third diagnostic process, all transistors 12 and 13 are placed in a passive state (and thus disconnected).

[0161] In the second step of the initial configuration of phase switches 7, all of these phase switches 7 are placed and kept in the closed state so that the phase u, v, w between the phase switches and the motor can be diagnosed.

[0162] In the phase biasing step, biasing module 5 controls biasing circuit 4 to bias all phases u, v, and w with a predetermined diagnostic voltage Vb. In this example, the diagnostic voltage Vb is 50% of the DCLink supply voltage of inverter 2.

[0163] In the voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w using voltage module 8. In this step, the average value of the three phase voltages is also calculated. The three measured voltages are added together, and the result is divided by three. This allows the voltage value generated by the back electromotive force to be obtained, thus independent of the back electromotive force changes caused by motor rotation.

[0164] In the comparison step, the average of the three phase voltages is compared with the expected result value. In this example, the expected result value is equal to the diagnostic voltage Vb.

[0165] Conversely, if the average of the three phase voltages is lower than voltage Vb (possibly minus a safety threshold), or if the average of the three phase voltages is lower than 0V plus a predetermined threshold, this indicates a short circuit between at least one phase and ground. Similarly, if the average of the three phase voltages is higher than voltage Vb (possibly plus a safety threshold), or if the average of the three phase voltages is higher than DCLink minus a predetermined threshold, this indicates a short circuit between at least one phase and the supply voltage DCLink.

[0166] At the end of this third diagnostic process, if one or more phases on the motor side are identified as short-circuited to ground or to the supply voltage, the system activates the indicator function as a sign of a short circuit in the motor (in Figure 2 Step E6).

[0167] If no short circuit is detected on any phase of the motor after the third sequential diagnostic process, the fourth sequential diagnostic process is then performed. Figure 2 Steps E5a and E5b). This fourth diagnostic procedure enables the identification of functional faults associated with one or more disconnected motor phases, namely, an interruption in the electrical conductor between the phase switch and the corresponding motor winding.

[0168] Depending on the electric speed, this fourth diagnostic process has two variations. The electric speed corresponds to the motor speed multiplied by a factor related to the number of pole pairs in the motor. For example, for a motor with four pole pairs, the ratio of electric speed to motor speed is 4. The first variation is implemented, for example, at electric speeds below 210 rad / s, while the second variation is implemented, for example, at high electric speeds above 210 rad / s. As variations, the speed threshold used to select the first or second variation can be dynamically adapted, especially based on the available supply voltage.

[0169] In its first variant (low motor speed), the fourth diagnostic process begins with the first step of the initial inverter configuration, whereby transistors 12 and 13, except for those in one branch of the inverter corresponding to a phase u, v, w, referred to as the phase being tested, are placed in a passive state (i.e., disconnected). The three phases u, v, w are tested sequentially using three iterations of this method, each iteration allowing testing of one phase u, v, w. For example, in the first iteration, the first phase to be tested is phase u. While the transistors in the other two phases v and w remain disconnected, the transistor corresponding to phase u is then controlled according to pulse width modulation (PWM) control, which is preferably selected to be far from 50%, i.e., having a high value (e.g., 80%) or a low value (e.g., 20%).

[0170] In the second step of the initial configuration of phase switches 7, all of these phase switches 7 are placed and kept in the closed state.

[0171] In the phase biasing step, biasing module 5 controls biasing circuit 4 to bias all phases u, v, and w with a predetermined diagnostic voltage Vb. In this example, the diagnostic voltage Vb is 50% of the DCLink supply voltage of inverter 2.

[0172] In the voltage measurement step, the microcontroller uses voltage module 8 to acquire the voltage of each phase u, v, w.

[0173] In the comparison step, the voltage measurements for each phase u, v, and w are compared with the expected results. The expected result for the phase being tested is equal to the supply voltage DCLink multiplied by the duty cycle. The expected results for the other two phases are equal to the preceding voltage (supply voltage DCLink multiplied by the duty cycle) plus the back electromotive force of the phase involved.

[0174] Conversely, if the phase being tested is open-circuited at the motor, the voltage measured on the phase being tested is equal to the supply voltage DCLink multiplied by the duty cycle, while the voltages measured on the other two phases are equal to the diagnostic voltage Vb plus the back electromotive force of the phase involved.

[0175] For example, if the test is performed on phase u:

[0176] - The following voltage values ​​correspond to the expected results and no faults were detected: the value for phase u is DCLink multiplied by the duty cycle; the value for phase v is DCLink multiplied by the duty cycle plus the back EMF of phase v; and the value for phase w is DCLink multiplied by the duty cycle plus the back EMF of phase w.

[0177] - The following voltage values ​​correspond to the functional failures identified as being related to the interruption of the phase u conductor on the motor side: the value for phase u is DCLink multiplied by the duty cycle; the value for phase v is Vb plus the back EMF of phase v; and the value for phase w is Vb plus the back EMF of phase w.

[0178] For example, the back electromotive force (EMF) of each phase can be calculated. In practice, the magnitude of the back EMF voltage generated by the motor speed is equal to the motor speed multiplied by the square root of 3 (motor flux). The motor speed is measured by a speed module. Alternatively, since this first variant is applied at low speeds, a threshold can be used to estimate the back EMF of each phase.

[0179] The three iterations of the first variant of the diagnosis (slow motor speed) enable each phase to switch sequentially to the role of the phase being tested, thereby enabling the identification of the aforementioned functional failures in all phases u, v, and w.

[0180] At the end of this first variation of the fourth diagnostic process, if one or more phases on the motor side are identified as open circuits, the system activates an indicator that the malfunction is an open circuit in the motor and indicates the flag of the affected phase (in...). Figure 2 Step E6).

[0181] In its second variant (fast motor speed), the fourth diagnostic process begins with the first step of the initial inverter configuration, where the step involves disabling the control of the power stage, i.e., during this second variant of the fourth diagnostic process, all transistors 12 and 13 are placed in a passive state (and thus disconnected).

[0182] In the second step of the initial configuration of phase switches 7, all of these phase switches 7 are placed and kept in the closed state.

[0183] In the phase biasing step, biasing module 5 controls biasing circuit 4 to bias all phases u, v, and w with a predetermined diagnostic voltage Vb. In this example, the diagnostic voltage Vb is 50% of the DCLink supply voltage of inverter 2.

[0184] In the voltage measurement step, the microcontroller acquires the voltage of each phase u, v, w using voltage module 8. In this example, N voltage values ​​are sampled simultaneously on each phase (e.g., 20). Therefore, sample "phase u(n)" represents the voltage measurement performed on phase u at time n, sample "phase v(n)" represents the voltage measurement performed on phase v at the same time n, and sample "phase w(n)" represents the voltage measurement performed on phase w at the same time n. For example, if two voltage measurement samples are taken on phases u, v, w (N=2), then two sets of three voltage measurement values ​​will be available:

[0185] - Three simultaneous measurements (n=1): phase u(1), phase v(1), phase w(1);

[0186] - Three simultaneous measurements (n=2): phase u(2), phase v(2), phase w(2).

[0187] In the step of measuring motor speed, the sampling of N speed values ​​is performed simultaneously with the voltage measurement in the previous step. Therefore, each speed measurement sample, denoted as MotSpdEl(n), corresponds to a motor speed measurement value acquired simultaneously with the voltage samples of phase u(n), phase v(n), and phase w(n).

[0188] In the comparison step, the voltage measurements for each phase u, v, w are compared with the expected result values. This comparison step performs a comparison between the representative voltage Vdiag determined for each phase u, v, w and the threshold voltage Vseuil determined as shown below. The principle is that the expected result value is:

[0189] - For phase u: Vb + the back electromotive force of phase u;

[0190] - For phase v: Vb + the back electromotive force of phase v;

[0191] - For phase w: Vb + the back electromotive force of phase w;

[0192] - The sum of the back electromotive force of phase u, the back electromotive force of phase v, and the back electromotive force of phase w is zero.

[0193] Therefore, if the test is performed on phase v according to the example above, and if the phase is open-circuited at the motor, the voltage measured on phase v will be approximately equal to the diagnostic voltage Vb, and the sum of the back EMF of phase u and the back EMF of phase w will be zero.

[0194] In practice, the representative voltage Vdiag for phases u, v, and w is preferably calculated for each measured voltage sample n. The representative voltage Vdiag for one phase (referred to as the phase under test) and for voltage measurement sample "n" is equal to the absolute value of the average voltage n of all phases u, v, and w minus the average voltage n of the two phases not under test. The calculation of this representative voltage Vdiag is written as:

[0195]

Mathematical Formula 1

[0196] .

[0197] Therefore, the value VdiagPhX(n) is calculated over three iterations for each phase u, v, w, in which phase u, v, w are sequentially assigned as the phase being tested.

[0198] Furthermore, the threshold voltage is calculated for each motor speed measurement sample "n" according to the following formula:

[0199]

Mathematical Formula 2

[0200]

[0201] In the above formula:

[0202] - VseuilMin is the set safety threshold, which is set to 0.3V in this example;

[0203] - SpdRatio is the ratio applied to the motor speed to increase the threshold voltage. In this example, SpdRatio is set to 0.0015V / (rad / s);

[0204] - MotSpdEl(n) is the measured motor speed of sample n;

[0205] - MotSpdElmin is the minimum speed that can be used for this second variant of the fourth diagnostic process. Here, the value is 210 rad / s.

[0206] A comparison step is performed by comparing the value VdiagPhX(n) and the corresponding value Vseuil(n) for each phase u, v, w. When VdiagPhX(n) is lower than Vseuil(n), an error is identified for sample n and for the involved phases u, v, w. When more than 80% of the samples n for a given phase u, v, w result in an error, a fault is identified, i.e., a discontinuity of the motor phase in the involved phase. The 80% ratio corresponds to an adjustable identification threshold.

[0207] At the end of this second variation of the fourth diagnostic process, if phases u, v, and w are identified as having an open circuit at the motor, the system activates a flag (in) by activating a signal indicating which phase is open. Figure 2 Step E6).

[0208] Figure 4 An example of bias module 5 is shown, which is adapted to generate a diagnostic voltage Vb equal to 50% of the supply voltage DCLink for each of phases u, v, and w. This central value allows for as much separation as possible from both the zero potential of the ground wire and the supply potential, enabling high-speed diagnostics regardless of the back EMF voltage caused by the motor speed.

[0209] The voltage divider circuit includes a logic signal input 14 connected to the microcontroller, which activates or deactivates phase bias. The logic signal input 14 controls a MOS transistor 15 positioned between the supply voltage DCLink and ground via a bipolar transistor 17. When the MOS transistor 15 is controlled (to bias the phases during the phase biasing step), the motor phases are connected to the supply voltage DCLink via so-called "pull-up" resistors 18. When the MOS transistor 15 is not controlled, phases u, v, and w are grounded via so-called "pull-down" resistors 19.

[0210] Resistors 18 and 19 use high resistance values ​​so that the circuit will not interfere with the power stage's control of the motor phase when the power stage is active.

[0211] This allows any other electronic circuit capable of applying the diagnostic voltage Vb to phases u, v, w to be used alternatively as bias module 5.

[0212] Other variations of the functional impairment identification method may be implemented without departing from the scope of the invention. For example, only some of the described sequential diagnostic procedures may be performed to constitute the functional impairment identification method.

Claims

1. A method of malfunction identification of an inverter-multiphase electric machine assembly, said inverter-multiphase electric machine assembly comprising an electric machine (1) and an inverter (2), said inverter (2) having power switches (12, 13) distributed on branches, each phase of said electric machine (1) being connected to one branch of the inverter (2) through one phase switch (7), the method being characterized in that it comprises a plurality of sequential diagnostic procedures, each sequential diagnostic procedure comprising the following steps: - an initial configuration step of the inverter (2) in which a control selection is made for each branch of the inverter, the selection being made from the group comprising: fixed control of the power switches (12, 13) in open mode; pulse width modulation control according to a predetermined diagnostic duty cycle of the power switches (12, 13); - an initial configuration step of the phase switches (7) in which a state selection is made for each phase switch (7), the selection being made between the closed state and the open state; - a phase (u, v, w) biasing step in which a predetermined diagnostic voltage (Vb) is applied to each branch of the inverter (2); - a voltage measurement step in which the voltage of each phase (u, v, w) is measured; - a comparison step in which the voltage measurement of each phase (u, v, w) is compared to an expected result value; - a malfunction identification step when the voltage measurement of one phase is different from the expected result value, and said method further comprising: a first sequential diagnostic procedure in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) are controlled in open mode; - in the initial configuration step of the phase switches (7), all the phase switches (7) are placed in the open state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage is 50% of the supply voltage (DC Link) of the inverter (2); - in the comparison step, the expected result value of each phase is equal to the predetermined diagnostic voltage (Vb); - in the malfunction identification step, when the voltage measurement of a phase is equal to the supply voltage (DC Link) of the inverter (2), a short circuit is identified on the high side power switch (12) of the corresponding phase, a second sequential diagnostic procedure in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) except the one corresponding to the phase being tested are controlled in open mode, the power switch corresponding to the phase being tested being pulse width modulated according to a predetermined diagnostic duty cycle; - in the initial configuration step of the phase switches (7), all the phase switches (7) are placed in the open state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage (Vb) is 50% of the supply voltage (DC Link) of the inverter (2); - in the comparison step, the expected result value of the phase being tested is a value equal to the supply voltage (DC Link) of the inverter (2) multiplied by the predetermined diagnostic duty cycle, and the expected result value of the other phases is equal to the predetermined diagnostic voltage (Vb), a third sequential diagnostic procedure in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) are controlled in open mode; - in the initial configuration step of the phase switches (7), all the phase switches (7) are put in the closed state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage (Vb) is 50% of the supply voltage (DCLink) of the inverter (2); - in the voltage measurement step, the average value of the voltages of all the phases (u, v, w) is determined; - in the comparison step, the expected result value of the average value of the voltages of all the phases is equal to the predetermined diagnostic voltage (Vb); - in the malfunction identification step, when the average value of the voltages of all the phases is lower than the predetermined diagnostic voltage (Vb), a short circuit between at least one phase and the ground is identified, the short circuit being located between the phase switch (7) and the motor (1).

2. The method of claim 1, wherein, In the phase biasing step, the predetermined diagnostic voltage is 50% of the supply voltage of the inverter (2).

3. The method of claim 1, wherein, The voltage measurement step is performed by determining the average value of a plurality of voltage measurement samples taken for each phase (u, v, w).

4. The method according to any one of claims 1 to 3, characterized in that, In the malfunction identification step, when the voltage measurement of a phase is equal to zero, a short circuit is identified on the low-side power switch (13) of the corresponding phase.

5. The method according to any one of claims 1 to 3, characterized in that, According to a first part of a second sequential diagnostic procedure: - in the initial configuration step of the inverter, the predetermined diagnostic duty cycle is lower than 50%; - in the malfunction identification step, when the voltage measurement of the phase being tested is equal to the voltage of another phase not being tested, a short circuit between these two phases is identified.

6. The method according to any one of claims 1 to 3, characterized in that, When the voltage measurement of the phase being tested is higher than the supply voltage (DCLink) multiplied by the predetermined diagnostic duty cycle, a malfunction is identified in the control of the low-side power switch (13) of the phase being tested.

7. The method of claim 5, wherein, According to a second part of the second sequential diagnostic procedure: - in the initial configuration step of the inverter (2), the predetermined diagnostic duty cycle is higher than 50%; - in the malfunction identification step, when the voltage measurement of the phase being tested is lower than the supply voltage (DCLink) multiplied by the predetermined diagnostic duty cycle, a malfunction is identified in the control of the high-side power switch (12) of the phase being tested.

8. The method according to any one of claims 1 to 3, characterized in that, In the malfunction identification step, when the average value of the voltages of all the phases (u, v, w) is higher than the predetermined diagnostic voltage (Vb), a short circuit between at least one phase and the supply voltage (DCLink) of the inverter (2) is identified, the short circuit being located between the phase switch (7) and the motor (1).

9. The method according to any one of claims 1 to 3, characterized in that, It comprises a first variant of the fourth sequential diagnostic procedure, in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) except the power switches corresponding to the phase being tested are controlled in open mode, the power switches corresponding to the phase being tested being controlled in pulse width modulation according to a predetermined diagnostic duty cycle; - in the initial configuration step of the phase switches (7), all the phase switches (7) are put in the closed state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage (Vb) is 50% of the supply voltage of the inverter (2); - in the voltage measurement step, the average value of the voltages of all the phases (u, v, w) is determined; - in the comparison step, the expected result value of the average value of the voltages of all the phases is equal to the predetermined diagnostic voltage (Vb); - in the malfunction identification step, when the average value of the voltages of all the phases is lower than the predetermined diagnostic voltage (Vb), a short circuit between at least one phase and the ground is identified, the short circuit being located between the phase switch (7) and the motor (1). - in the comparison step, the expected result value for the phase being tested is equal to the value obtained by multiplying the supply voltage of the inverter (2) (DC Link) by a predetermined diagnostic duty cycle, and the expected result values for the other phases are equal to the voltage of the phase being tested plus the back electromotive force of the phase concerned; - in the malfunction identification step, a short-circuit is identified between the phase switch (7) and the motor (1) on the first phase when the voltage of the first phase is equal to the value obtained by multiplying the supply voltage of the inverter (2) (DC Link) by a predetermined diagnostic duty cycle, and when the sum of the back electromotive forces of the other phases is equal to zero.

10. The method of claim 9, wherein, It comprises a second variant of the fourth sequential diagnostic process, in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) are controlled in open mode; - in the initial configuration step of the phase switches (7), all the phase switches (7) are placed in the closed state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage (Vb) is 50% of the supply voltage of the inverter (2) (DC Link); - in the comparison step, the expected result value for each phase is equal to the predetermined diagnostic voltage (Vb) plus the back electromotive force of the phase concerned, where the sum of the back electromotive forces of all the phases is equal to zero; - in the malfunction identification step, a short-circuit is identified between the phase switch (7) and the motor (1) on the first phase when the voltage of the first phase is equal to the predetermined diagnostic voltage (Vb) and when the sum of the back electromotive forces of the other phases is equal to zero.

11. The method of claim 10, wherein, It comprises a second variant of the fourth sequential diagnostic process, in which: - in the initial configuration step of the inverter (2), all the power switches (12, 13) are controlled in open mode; - in the initial configuration step of the phase switches (7), all the phase switches (7) are placed in the closed state; - in the phase (u, v, w) biasing step, the predetermined diagnostic voltage (Vb) is 50% of the supply voltage of the inverter (2) (DC Link); - in the measurement step, the voltage measurements are performed from a sample comprising a predetermined number of measurements; - in the comparison step, the expected result value for each phase is equal to the predetermined diagnostic voltage (Vb) plus the back electromotive force of the phase concerned, where the sum of the back electromotive forces of all the phases is equal to zero; - in the malfunction identification step, an error is identified for the first phase when the voltage of the first phase is equal to the predetermined diagnostic voltage (Vb) and when the sum of the back electromotive forces of the other phases is equal to zero; - when the number of identified errors exceeds an identification threshold, a short-circuit is identified between the phase switch (7) and the motor (1) on the first phase.

12. The method according to one of claims 10 or 11, characterized in that, The first variant of the fourth sequential diagnostic process is implemented at electrical speeds below 200 rad / s, and the second variant of the fourth sequential diagnostic process is implemented at electrical speeds above 200 rad / s.

Images