A composite fault diagnosis method for distributed electric vehicle drive motors

Abstract

The present invention discloses a composite fault diagnosis method for distributed electric vehicle drive motors, which specifically relates to the technical field of electric vehicle fault diagnosis, including diagnosing the vehicle state, calculating the ideal yaw angular velocity ω based on the steering wheel angle signal σ and the accelerator pedal opening signal P input by the driver, and using a two-degree-of-freedom dynamics model d and the ideal center of mass sideslip angle β d , and the actual vehicle yaw rate ω measured by the sensor r and the actual center of mass sideslip angle β r The yaw rate error e is obtained by subtracting ω and the center of mass sideslip angle error e β , e ω and e β Respectively with their respective thresholds η ω and η β For comparison, if e ω and e β At the same time, it is less than or equal to the threshold value η ω and η β , the vehicle state is judged to be normal, otherwise the vehicle state is judged to be abnormal; when the vehicle state is abnormal, the drive motor PMSM diagnosis method based on the structural analysis method and the drive motor PMSM torque diagnosis method are started to diagnose and isolate the hub motor fault.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle fault diagnosis, and in particular to a composite fault diagnosis method for distributed electric vehicle drive motors. Background Art

[0002] Electric drive control is a key technical part of electric vehicles, and the drive motor is the core component of electric drive control.

[0003] Permanent magnet synchronous motor (PMSM) has become the first choice for most electric vehicle motors due to its high efficiency, large starting torque, low temperature rise, and small size. However, during vehicle driving, the motor's working environment is complex. When the permanent magnet material of PMSM is subjected to vibration, high temperature, and overload current, its magnetic conductivity may decrease or demagnetization may occur, which may reduce the performance of the permanent magnet motor. In addition, the control of PMSM requires multiple sensor signals, such as rotor position signals, current signals, etc. When any part of them fails, it will affect the drive motor's response to the torque signal. Therefore, it is necessary to perform fault detection on the drive motor.

[0004] Fault detection is the basis of active fault-tolerant control of vehicles. Obtaining vehicle fault information in a timely and accurate manner is of great significance to vehicle safety control.

[0005] Most fault detection methods currently only detect the fault once, which may lead to misdiagnosis. In addition, most fault diagnosis methods do not consider factors such as the fault form, location, and degree at the same time, and the fault information obtained is inaccurate or incomplete. If this incomplete or inaccurate information is provided to the controller, it may cause the controller to issue an erroneous command, thus threatening the driving safety of the vehicle. Summary of the invention

[0006] To this end, the present invention provides a composite fault diagnosis method for a distributed electric vehicle drive motor to solve the problems raised in the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a composite fault diagnosis method for a distributed electric vehicle drive motor, comprising:

[0008] S1: Diagnose the vehicle status and calculate the ideal yaw rate ω based on the steering wheel angle signal σ and accelerator pedal opening signal P input by the driver using a two-degree-of-freedom dynamics model d and the ideal center of mass sideslip angle β d , and the actual vehicle yaw rate ω measured by the sensor r and the actual center of mass sideslip angle β r The yaw rate error e is obtained by subtracting ωand the center of mass sideslip angle error e β , e ω and e β Respectively with their respective thresholds η ω and η β For comparison, if e ω and e β At the same time, it is less than or equal to the threshold value η ω and η β , the vehicle status is judged to be normal, otherwise the vehicle status is judged to be abnormal;

[0009] Among them, the yaw angular velocity error e obtained by using the two-degree-of-freedom dynamic model is ω and the center of mass sideslip angle error e β It is also possible not to send it to the vehicle controller. The reason for calculating the ideal state of the vehicle is to compare it with the actual state of the vehicle. This comparison is achieved by comparing the residual with the threshold. This process is essentially a process of calculation and logical judgment, which can be performed in the vehicle controller or in other vehicle-mounted MCUs.

[0010] S2: When the vehicle is in an abnormal state, the PMSM diagnosis method based on the structural analysis method and the PMSM torque diagnosis method of the drive motor are started to diagnose and isolate the hub motor fault. The former can obtain the fault form of the permanent magnet synchronous motor and the location of the faulty motor, and the latter can obtain the location of the faulty motor and the degree of the fault.

[0011] Preferably, the ideal yaw rate ω is obtained by using a two-degree-of-freedom dynamics model according to the steering wheel angle signal σ and the accelerator pedal opening signal P input by the driver. d and the ideal center of mass sideslip angle β d as follows:

[0012]

[0013]

[0014] Where: d represents the ideal yaw rate, β d represents the ideal center of mass sideslip angle, v x represents the longitudinal speed of the vehicle; m represents the total mass of the vehicle, L represents the wheelbase of the vehicle, l f , l r Respectively represent the front and rear wheelbase of the vehicle; C f , C r Represent the cornering stiffness of the front and rear wheels respectively; is the vehicle stability coefficient; σ represents the front wheel steering angle, the front wheel steering is adopted by default, and the rear wheel steering angle is zero; μ represents the road friction coefficient, g represents the gravity acceleration, sgn(·) represents the sign function, and ω represents the vehicle yaw rate, where the pedal depth input by the driver mainly affects the vehicle longitudinal speed v x .

[0015] Preferably, the steps of the PMSM diagnostic method for drive motors based on structural analysis are as follows: first, define the key faults of the PMSM system and establish a motor model containing the faults; with the help of the DM decomposition algorithm, evaluate the fault detectability and fault isolability of the PMSM system and obtain the minimum overdetermined equation set MSO sets; combine the minimum overdetermined equation set MSOsets and implement the residual design of the system based on the analytical redundancy relationship to obtain the specific fault form of the PMSM. In addition, since each hub motor has an independent sensor, structural analysis is performed on each drive motor to obtain the fault form of each motor and locate each faulty motor. Here, an arbitrary motor is taken as an example to establish a PMSM model.

[0016] The following model is established for PMSM:

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027] e 11 :V veh =W mh ·R wh ;

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037] e 21 :y W =W+f W ;

[0038] where v a ,v b ,v c Represents the three-phase voltage, i a ,i b ,i c is the three-phase current, v d ,v q Represent the d-axis and q-axis voltages respectively, i d ,i q Represent the d-axis and q-axis currents, L d ,L q Represent the d-axis and q-axis inductances respectively. m represents the rotor flux, T e Represents the motor electromagnetic torque, W m represents the motor speed, J represents the total moment of inertia of the motor rotating elements, θ e represents the electrical angle position, p is the number of motor pole pairs, and b represents the motor friction coefficient. L represents the load torque, the vehicle speed V veh It is related to the road slope. W represents the wheel speed. R s represents the stator phase resistance, i gear represents the transmission ratio from the motor to the wheel, m represents the total mass of the vehicle, g represents the acceleration of gravity, θ represents the angle between the road and the horizontal plane, μ represents the road friction coefficient, and C d is the air resistance coefficient, ρ A is the air density, A f is the frontal area of ​​the vehicle, V veh Represents the relative speed between the vehicle and the air. 13 ~e 21 An equation representing sensor measurement information including fault information. Indicates that the three-phase voltage sensor is faulty. Indicates that the three-phase current sensor is faulty. Indicates that the motor speed sensor is faulty. Indicates that the motor electrical angle position sensor is faulty, f W Indicates a wheel speed sensor malfunction. Indicates the measured value of the three-phase voltage sensor, Represents the three-phase current sensor measurement value, y Wm Represents the measured value of the motor speed sensor, y θe Represents the measured value of the motor electrical angle position sensor, y W Indicates the wheel speed sensor measurement value.

[0039] Preferably, the drive motor PMSM torque diagnosis method is to compare the expected torque of each hub motor calculated by the controller according to the driver input and the vehicle state error with the actual torque of each hub motor response of the vehicle to obtain the faulty motor position and fault degree.

[0040] Preferably, the fault information of S1 and S2 is integrated to output a diagnosis result, which includes the form, location and degree of the hub motor fault.

[0041] The present invention has the following advantages:

[0042] The present invention firstly performs multiple diagnosis on the drive motor of the distributed drive electric vehicle to ensure the accuracy of the diagnosis result; secondly, a method combining vehicle state judgment, structural analysis and torque diagnosis is used to diagnose the drive motor, and the diagnosis information is integrated to simultaneously know the fault type, fault form and fault degree of the drive motor, and the diagnosis information is accurate and complete. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The present invention provides ω 、e β Not inferior to the overall block diagram of vehicle controller fault diagnosis;

[0044] Figure 2 The present invention provides ω 、e β Input to the overall block diagram of vehicle controller fault diagnosis;

[0045] Figure 3 A structural analysis method fault diagnosis flow chart for PMSM provided by the present invention;

[0046] Figure 4 The PMSM structure analysis diagram provided by the present invention;

[0047] Figure 5The DM decomposition diagram of the PMSM model provided by the present invention;

[0048] Figure 6 This is a calculation flow chart of the minimum structure overdetermined equation set provided by the present invention. DETAILED DESCRIPTION

[0049] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Refer to the instruction manual Figure 1 , a composite fault diagnosis method for a distributed electric vehicle drive motor of this embodiment includes:

[0051] S1: Diagnose the vehicle status and calculate the ideal yaw rate ω based on the steering wheel angle signal σ and accelerator pedal opening signal P input by the driver using a two-degree-of-freedom dynamics model d and the ideal center of mass sideslip angle β d , and the actual vehicle yaw rate ω measured by the sensor r and the actual center of mass sideslip angle β r The yaw rate error e is obtained by subtracting ω and the center of mass sideslip angle error e β , e ω and e β Respectively with their respective thresholds η ω and η β For comparison, if e ω and e β At the same time, it is less than or equal to the threshold value η ω and η β , the vehicle status is judged to be normal, otherwise the vehicle status is judged to be abnormal;

[0052] Among them, the yaw angular velocity error e obtained by using the two-degree-of-freedom dynamic model is ω and the center of mass sideslip angle error e β It is also possible not to send it to the vehicle controller. The reason for calculating the ideal state of the vehicle is to compare it with the actual state of the vehicle. This comparison is achieved by comparing the residual with the threshold. This process is essentially a process of calculation and logical judgment, which can be performed in the vehicle controller or in other vehicle-mounted MCUs, such as Figure 1 and Figure 2 shown.

[0053] S2: When the vehicle is in an abnormal state, the PMSM diagnosis method based on the structural analysis method and the PMSM torque diagnosis method of the drive motor are started to diagnose and isolate the hub motor fault. The former can obtain the fault form of the permanent magnet synchronous motor and the location of the faulty motor, and the latter can obtain the location of the faulty motor and the degree of the fault.

[0054] Preferably, the ideal yaw rate ω is obtained by using a two-degree-of-freedom dynamics model according to the steering wheel angle signal σ and the accelerator pedal opening signal P input by the driver. d and the ideal center of mass sideslip angle β d as follows:

[0055]

[0056]

[0057] Where: d represents the ideal yaw rate, β d represents the ideal center of mass sideslip angle, v x represents the longitudinal speed of the vehicle; m represents the total mass of the vehicle, L represents the wheelbase of the vehicle, l f , l r Respectively represent the front and rear wheelbase of the vehicle; C f , C r Represent the cornering stiffness of the front and rear wheels respectively; is the vehicle stability coefficient; σ represents the front wheel steering angle, the front wheel steering is adopted by default, and the rear wheel steering angle is zero; μ represents the road friction coefficient, g represents the gravity acceleration, sgn(·) represents the sign function, and ω represents the vehicle yaw rate, where the pedal depth input by the driver mainly affects the vehicle longitudinal speed v x .

[0058] like Figure 1 As shown, during the working process, the ideal yaw rate and the ideal sideslip angle of the vehicle center of mass can be calculated based on the information measured by the vehicle sensor, and the difference between them and the actual yaw rate measured by the vehicle sensor is as follows:

[0059] e ω =ω d -ω r ,e β =β d -β r ;

[0060] where e ω represents the yaw rate error, ω d represents the desired yaw rate, ω r represents the actual yaw angular velocity; e β represents the sideslip angle error of the center of mass, β drepresents the ideal center of mass sideslip angle, β r Represents the actual center of mass sideslip angle.

[0061] The error is compared with the preset threshold to determine whether the vehicle status is normal, as shown in Table 1:

[0062] Table 1

[0063]

[0064] Among them, η ω represents the yaw rate error threshold, η β Represents the center of mass sideslip angle error threshold.

[0065] like Figure 3 As shown in the figure, the steps of the PMSM diagnosis method for drive motors based on structural analysis are as follows: first, define the key faults of the PMSM system and establish a motor model containing the faults; with the help of the DM decomposition algorithm, evaluate the fault detectability and fault isolability of the PMSM system, and obtain the minimum overdetermined equation set MSO sets; combine the minimum overdetermined equation set MSO sets, and based on the analytical redundancy relationship, implement the residual design of the system to obtain the specific fault form of the PMSM. In addition, since each wheel hub motor has an independent sensor, the structural analysis method is performed on each drive motor to obtain the fault form of each motor and locate each faulty motor. Here, an arbitrary motor is taken as an example to establish a PMSM model.

[0066] The PMSM model including fault information is established as follows:

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077] e 11 :V veh =W mh ·R wh ;

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087] e 21 :y W =W+f W ;

[0088] where v a ,v b ,v c Represents the three-phase voltage, i a ,i b ,i c is the three-phase current, v d ,v q Represent the d-axis and q-axis voltages respectively, i d ,i q Represent the d-axis and q-axis currents, L d ,L q Represent the d-axis and q-axis inductances respectively. m represents the rotor flux, T e Represents the motor electromagnetic torque, W m represents the motor speed, J represents the total moment of inertia of the motor rotating elements, θ e represents the electrical angle position, p is the number of motor pole pairs, and b represents the motor friction coefficient. L represents the load torque, the vehicle speed V veh It is related to the road slope. W represents the wheel speed. R s represents the stator phase resistance, i gearrepresents the transmission ratio from the motor to the wheel, m represents the total mass of the vehicle, g represents the acceleration of gravity, θ represents the angle between the road and the horizontal plane, μ represents the road friction coefficient, and C d is the air resistance coefficient, ρ A is the air density, A f is the frontal area of ​​the vehicle, V veh Represents the relative speed between the vehicle and the air. 13 ~e 21 An equation representing sensor measurement information including fault information. Indicates that the three-phase voltage sensor is faulty. Indicates that the three-phase current sensor is faulty. Indicates that the motor speed sensor is faulty. Indicates that the motor electrical angle position sensor is faulty, f W Indicates a wheel speed sensor malfunction. Indicates the measured value of the three-phase voltage sensor, Represents the three-phase current sensor measurement value, Represents the measured value of the motor speed sensor, y θe Represents the measured value of the motor electrical angle position sensor, y W Indicates the wheel speed sensor measurement value.

[0089] Define the variable measured by the sensor as a known variable, as shown in formula e 13 ~e 21 , then the known variables are

[0090] The unknown variable is {v a ,v b ,v c ,v d ,v q ,θ e ,i a ,i b ,i c ,i q ,i d ,W m ,T e ,T L ,V veh ,W};

[0091] The fault variable is

[0092] like Figure 4 As shown, the structural analysis diagram of PMSM can be established.

[0093] The present invention uses DM decomposition (Dulmage-Mendelsohn Decomposition) to analyze the redundancy of the system model. With the help of this mathematical tool, the structure matrix of the PMSM can be rearranged to be close to the upper right triangular matrix, and the PMSM system model can be structurally divided into three parts:

[0094] The structural underdetermined part M - : The number of equations in the equation set is less than the number of unknowns.

[0095] Structural positive part M 0 : The number of equations in the equation set is equal to the number of unknowns.

[0096] Structural overdetermined part M + : The number of equations in the equation set is greater than the number of unknowns.

[0097] When the fault is in the overdetermined part of the structure M + When the fault is detected, the PMSM model is generally a redundant system, so all the faults mentioned in the model can be detected.

[0098] According to the theory of structural analysis, when the fault f i In addition to the fault f j When the structure of the equation is overdetermined, it is proved that the fault f i Can and fault f j Isolation, that is:

[0099] Perform DM decomposition on the PMSM model, and the decomposition results are as follows: Figure 5 shown.

[0100] When the redundancy of an overdetermined set of equations is exactly equal to 1, the overdetermined set of equations becomes a minimum structured overdetermined set (MSOsets). The purpose of solving the minimum structured overdetermined set is to use as few equations as possible to detect a specific fault. The solution process of MSO sets is as follows: Figure 6 shown.

[0101] In the PMSM model, multiple minimum structured supersets can be constructed through DM decomposition to isolate all sensor faults. Assuming that only one sensor fails each time, the following nine minimum structured supersets are selected. The equations contained in the selected minimum structured supersets are shown in Table 2:

[0102] Table 2

[0103]

[0104] The fault variables included in each minimum structural overdetermined equation set are shown in Table 3:

[0105] Table 3

[0106]

[0107]

[0108] “√” in the table indicates that the fault variable is included in the minimum structural overdetermined equation set.

[0109] After obtaining the minimum structured overdetermined equation set, it is necessary to generate the residual of each minimum structured overdetermined equation set. The residual generation method includes a sequence residual generation method, etc., which will not be specifically described in the present invention.

[0110] After obtaining the residuals of each minimum structure overdetermined equation set, the PMSM fault detection can be performed. The type of fault in the motor can be determined based on the change in the residual value. For example, if the PMSM electrical angle position sensor fails, When the residual value of MSO2-MSO9 changes and exceeds the given threshold, the residual value change of MSO1 will not exceed the threshold. At this time, it can be determined that the motor electrical angle position sensor is faulty. Similarly, when the wheel speed sensor fails, that is, f W ≠0, the residuals of MSO1 and MSO2 will change and exceed the given threshold, while the changes of MSO3-MSO9 will not exceed the threshold. At this time, it can be determined that the motor electrical angle position sensor is faulty. The other fault detection methods are similar to the above. When none of the nine faults mentioned in the modeling occurs, the residual values ​​of MOS1-MSO9 will not change or the changes will be less than the given threshold. When a PMSM fault occurs, after performing fault detection on the PMSM using the structural analysis method, the fault type of the PMSM can be known.

[0111] The above-mentioned PMSM diagnostic method for drive motor based on structural analysis method detects a single PMSM, while the vehicle type studied in the present invention is a distributed drive electric vehicle with four wheel hub motors, that is, four PMSMs. The PMSM diagnostic method for drive motor based on structural analysis method involved in the present invention is applicable to each drive motor, so the fault analysis method can not only obtain the fault type of PMSM, but also locate the faulty motor, that is, it can detect the position of the faulty motor. Based on the PMSM diagnostic method for drive motor based on structural analysis method, the obtained fault information includes the position of the faulty motor and the fault type of the faulty motor.

[0112] When actively fault-tolerant control is performed on a vehicle, it is necessary not only to know the fault type and location of the faulty motor, but also to know the fault degree of the faulty motor. In actual application, the torque response of the drive motor is an important indicator for judging the degree of motor failure. The drive motor PMSM torque diagnosis method in the present invention can detect the response degree of each drive motor. The drive motor PMSM torque diagnosis method is to compare the ideal torque of each wheel drive motor calculated by the controller based on the driver input and the vehicle state error with the actual corresponding torque of each drive motor of the vehicle, so as to obtain the fault location and fault degree of the faulty motor, as follows:

[0113] Definition Ti =T ir -T id is the torque response error, where T id Indicates the ideal torque of each wheel drive motor sent by the controller, T ir Indicates the actual response torque of each wheel drive motor. For the subscripts id and ir, i = (fl, fr, rl, rl), representing the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively.

[0114] Torque Failure Factor T low Represents the smaller value of the absolute value of the torque, in order to prevent the appropriate T id → ΔT at 0 i →∞ situation.

[0115] When |T id |>T low When the threshold η is set T , if ΔT i ≤η T , then it is considered that the i-th drive motor has no fault, otherwise it has a fault;

[0116] When |T id |≤T low When the threshold τ is set T , if ΔT i ≤τ T , then it is considered that the i-th drive motor has no fault, otherwise it has a fault; since each drive motor has an independent torque sensor, the position of the faulty motor can be located.

[0117] Defining the torque fault level The fault degree of the motor can be obtained.

[0118] Finally, the acquired fault information is integrated to output the fault diagnosis results: the drive motor fault type, location and severity.

[0119] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A composite fault diagnosis method for distributed electric vehicle drive motors, characterized in that: include: S1: Diagnose the vehicle status, send the steering wheel angle signal σ and the accelerator pedal opening signal P input by the driver to the vehicle controller and the vehicle reference model at the same time, and use the two-degree-of-freedom dynamics model to calculate the ideal yaw angular velocity ω d and the ideal center of mass sideslip angle β d , and the actual vehicle yaw rate ω measured by the sensor r and the actual center of mass sideslip angle β r The yaw rate error e is obtained by subtracting ω and the center of mass sideslip angle error e β , e ω and e β Respectively with their respective thresholds η ω and η β For comparison, if e ω and e β At the same time, it is less than or equal to the threshold η ω and η β , the vehicle status is judged to be normal, otherwise the vehicle status is judged to be abnormal; S2: When the vehicle is in an abnormal state, the PMSM diagnosis method based on the structural analysis method and the PMSM torque diagnosis method based on the drive motor are started to diagnose and isolate the hub motor fault. The former obtains the fault form of the permanent magnet synchronous motor and the location of the faulty motor, and the latter obtains the location of the faulty motor and the degree of the fault. The steps of the PMSM diagnosis method for drive motors based on structural analysis are as follows: first, define the key faults of the PMSM system and establish a motor model containing the faults; use the DM decomposition algorithm to evaluate the fault detectability and fault isolation of the PMSM system and obtain the minimum overdetermined equation set MSO sets; combine the minimum overdetermined equation set MSO sets and implement the residual design of the system based on the analytical redundancy relationship to obtain the specific fault form of the PMSM. In addition, since each wheel hub motor has an independent sensor, the structural analysis method is used to detect each drive motor to obtain the fault form of each motor, and complete the positioning of each faulty motor to establish a PMSM model containing fault information, as shown below: e 11 :V veh =W mh ·R wh ; e 21 :y W =W+f W ; where v a ,v b ,v c Represents the three-phase voltage, i a ,i b ,i c is the three-phase current, v d ,v q Represent the d-axis and q-axis voltages respectively, i d ,i q Represent the d-axis and q-axis currents, L d ,L q Represents the d-axis and q-axis inductances respectively; λ m represents the rotor flux, T e Represents the motor electromagnetic torque, W m represents the motor speed, J represents the total moment of inertia of the motor rotating elements, θ e represents the electrical angle position, p is the number of motor pole pairs, b represents the motor friction coefficient; T L represents the load torque, the vehicle speed V veh It is related to the road slope; W represents the wheel speed; R s represents the stator phase resistance, i gear represents the transmission ratio from the motor to the wheel, m represents the total mass of the vehicle, g represents the acceleration of gravity, θ represents the angle between the road and the horizontal plane, μ represents the road friction coefficient, and C d is the air resistance coefficient, ρ A is the air density, A f is the frontal area of ​​the vehicle, V veh represents the relative speed between the vehicle and the air; e 13 ~e 21 An equation representing sensor measurement information including fault information; Indicates that the three-phase voltage sensor is faulty. Indicates a three-phase current sensor failure, f Wm Indicates that the motor speed sensor is faulty, f θe Indicates that the motor electrical angle position sensor is faulty, f W Indicates wheel speed sensor failure; Indicates the measured value of the three-phase voltage sensor, Represents the three-phase current sensor measurement value, Indicates the measured value of the motor speed sensor. Represents the measured value of the motor electrical angle position sensor, y W Indicates the wheel speed sensor measurement value; Define the variable measured by the sensor as a known variable, as shown in formula e 13 ~e 21 , then the known variables are The unknown variables are {v a , v b , v c , v d , v q , θ e , i a , i b , i c , i q , i d , W m , T e , T L , V veh , W}; The fault variable is The redundancy of the system model is analyzed with the help of DM decomposition. With the help of this mathematical tool, the PMSM structure matrix is ​​rearranged to be close to the upper right triangular matrix, and the PMSM system model is structurally divided into three parts: The structural underdetermined part M - : The number of equations in the equation set is less than the number of unknowns; Structural positive part M 0 : The number of equations in the equation set is equal to the number of unknowns; Structural overdetermined part M + : The number of equations in the equation set is greater than the number of unknowns; When the fault is in the overdetermined part of the structure M + When the fault is detected, the PMSM model is generally a redundant system, so all faults mentioned in the model can be detected. According to the theory of structural analysis, when the fault f i In addition to the fault f j When the structure of the equation is overdetermined, it is proved that the fault f i Can and fault f j Isolation, that is: Perform DM decomposition on the PMSM model; When the redundancy of the overdetermined equation set is exactly equal to 1, the overdetermined equation set becomes the minimum structured overdetermined set MSOsets. The purpose of solving the minimum structured overdetermined set is to use as few equations as possible to detect specific faults. In the PMSM model, multiple minimum structured supersets can be constructed through DM decomposition to isolate all sensor faults. Assuming that only one sensor fails each time, the following nine minimum structured supersets are selected. The equations contained in the selected minimum structured supersets are shown in Table 1 below: Table 1 The fault variables included in each minimum structural overdetermined equation set are shown in Table 2: Table 2 "√" in the table indicates that the fault variable is included in the minimum structural overdetermined equation set; After obtaining the minimum structure overdetermined equation set, it is necessary to generate the residual of each minimum structure overdetermined equation set. The residual generation method includes the sequence residual generation method; After obtaining the residuals of each minimum structural overdetermined equation set, PMSM is detected for faults; the type of fault in the motor is determined based on the change in the residual value; when none of the nine faults mentioned in the modeling occurs, the residual values ​​of MOS1-MSO9 do not change or the change is less than the given threshold; when a PMSM fault occurs, the PMSM fault type is known after performing fault detection on the PMSM using the structural analysis method; The above-mentioned PMSM diagnosis method for drive motor based on structural analysis is for detection of a single PMSM, while the distributed drive electric vehicle has four wheel hub motors, that is, four PMSMs. The PMSM diagnosis method for drive motor based on structural analysis is applicable to each drive motor, so the fault analysis method can not only know the fault type of PMSM, but also locate the faulty motor, that is, can detect the position of the faulty motor. Based on the PMSM diagnosis method for drive motor based on structural analysis, the fault information obtained includes the position of the faulty motor and the fault type of the faulty motor. When actively fault-tolerant control is performed on a vehicle, it is necessary not only to know the fault type and location of the faulty motor, but also the fault degree of the faulty motor. In actual application, the torque response of the drive motor is an important indicator for judging the fault degree of the motor. The PMSM torque diagnosis method of the drive motor can detect the response degree of each drive motor. The drive motor PMSM torque diagnosis method is to compare the expected torque of each wheel hub motor calculated by the controller according to the driver input and vehicle state error with the actual torque of each wheel hub motor response of the vehicle to obtain the fault motor position and fault degree; The details are as follows: Definition Ti =T ir -T id is the torque response error, where T id Indicates the ideal torque of each wheel drive motor sent by the controller, T ir Indicates the actual response torque of each wheel drive motor; for subscripts id and ir, i = (fl, fr, rl, rl), representing the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively; Torque Failure Factor T low Represents a smaller value of the absolute value of the torque, in order to prevent id → ΔT at 0 i →∞ case; When T id >T low When the threshold η is set T , if ΔT i ≤η T , then it is considered that the i-th drive motor has no fault, otherwise it has a fault; When T id ≤T low When the threshold τ is set T , if ΔT i ≤τ T , then it is considered that the i-th drive motor has no fault, otherwise it has a fault; since each drive motor has an independent torque sensor, it is convenient to locate the position of the faulty motor; Defining the torque fault level Get the fault degree of the motor.

2. A composite fault diagnosis method for distributed electric vehicle drive motors according to claim 1, characterized in that: According to the steering wheel angle signal σ and the accelerator pedal opening signal P input by the driver, the ideal yaw angular velocity ω is calculated with the help of the two-degree-of-freedom dynamic model. d and the ideal center of mass sideslip angle β d as follows: Where: d represents the ideal yaw rate, β d represents the ideal center of mass sideslip angle, v x represents the longitudinal speed of the vehicle; m represents the total mass of the vehicle, L represents the wheelbase of the vehicle, l f , l r Respectively represent the front and rear wheelbase of the vehicle; C f , C r Represent the cornering stiffness of the front and rear wheels respectively; is the vehicle stability coefficient; σ represents the front wheel steering angle, μ represents the road friction coefficient, g represents the gravitational acceleration, sgn(·) represents the sign function, and ω represents the vehicle yaw rate.

3. The composite fault diagnosis method for distributed electric vehicle drive motor according to claim 1 is characterized in that: The fault information of S1 and S2 is integrated to output the diagnosis result, which includes the form, location and degree of the hub motor fault.

Images