Fault diagnosis and fault-tolerant control method and system for duplex-winding permanent magnet synchronous motor

Abstract

The invention provides a fault diagnosis and fault-tolerant control method and system for a duplex-winding permanent magnet synchronous motor, and the method comprises the steps: designing a diagnosis method for an open-circuit fault, a grounding fault and a turn-to-turn short circuit fault for a target motor of which the type is the duplex-winding permanent magnet synchronous motor, carrying out the fault diagnosis of the target motor, and outputting a motor fault state mark; distinguishing corresponding motor fault states according to the obtained motor fault state marks; if the target motor has no fault, corner tracking control is carried out on the target motor; if the target motor is in the fault state, triggering corresponding fault-tolerant control schemes according to different fault states and applying corresponding constraint conditions; if the state is the healthy state, the control current is only determined by the motor model of the target motor and the rotation angle tracking controller, and if the state is the fault state, additional constraint is applied to the control current according to the fault type on the basis of the healthy state, so that the double-winding permanent magnet synchronous motor can stably operate in both the healthy state and the fault state.

Description

Technical Field

[0001] This invention proposes a fault diagnosis and fault-tolerant control method and system for a dual-winding permanent magnet synchronous motor, which relates to the field of intelligent connected vehicle drive-by-wire chassis technology. Background Technology

[0002] With the development trend of automotive intelligence and autonomous driving, steering-by-wire (SBW) systems, by eliminating the physical connection between the steering wheel and steering gear, offer significant advantages in design flexibility, transmission ratio adjustment, and path planning adaptability. Dual-winding permanent magnet synchronous motors, with their hardware redundancy, have become the core actuators meeting the highest safety requirements of ISO 26262 standard ASIL D. However, in practical applications, fault diagnosis and fault-tolerant control of this motor still face many pressing technical challenges, severely impacting the safety and steering angle tracking accuracy of the SBW system.

[0003] For the three common faults of dual-winding permanent magnet synchronous motors—single-phase open circuit, winding grounding, and inter-turn short circuit—existing methods have significant shortcomings in diagnosis: 1) Although single-phase open circuit manifests as a sudden drop in current, it is easily confused with normal current fluctuations under load fluctuation scenarios, and existing detection methods are prone to missed detection; 2) Winding grounding disrupts the three-phase current balance, and the abnormal fluctuation characteristics of zero-sequence current are easily masked by electromagnetic interference, making it difficult for traditional threshold diagnosis to accurately capture; 3) Inter-turn short circuits initially involve only a few coils, and the changes in external electrical characteristics are weak, making them easily masked by sensor noise and operating condition fluctuations. Existing analytical model methods cannot achieve a balance between sensitivity and robustness, and learning-based diagnostic methods face a contradiction between computing power and accuracy—high-precision algorithms such as deep learning require large-scale data processing, which is incompatible with the limited computing power of onboard MCUs, making it difficult to achieve lightweight deployment. At the same time, under dynamic operating conditions such as vehicle start-stop and large-angle steering, the transient error of the current loop is highly similar to the fault residual, and fixed threshold logic is prone to false alarms, leading to unnecessary system degradation and affecting driving continuity.

[0004] More critically, existing technologies have not yet formed a complete fault-tolerant control system, resulting in a severe deficiency in safety assurance capabilities after a fault occurs. Specifically: 1) When a motor experiences a single-phase open circuit, winding grounding, or inter-turn short circuit fault, the lack of targeted fault isolation strategies and topology reconfiguration schemes makes it impossible to quickly cut off the fault source, leading to a high risk of fault propagation; 2) The lack of current regulation and torque distribution mechanisms adapted to fault conditions makes it easy for motor torque output to become uncontrollable after a fault occurs, and the steering system cannot maintain basic angle tracking capability; 3) Even if some solutions attempt fault-tolerant processing, they generally do not consider torque ripple compensation and coordinated steering stability control, resulting in a significant decrease in driving smoothness and steering accuracy under fault conditions, even failing to meet the basic requirements for safe driving. In summary, existing technologies, when dealing with the three types of frequently occurring faults, suffer from insufficient diagnostic accuracy, difficulty in balancing computing power and diagnostic performance, and a tendency to generate false alarms under complex operating conditions. Furthermore, they lack a comprehensive fault-tolerant solution covering the entire process from fault isolation to torque control, and from topology reconfiguration to angle tracking. Therefore, there is an urgent need to develop a systematic fault diagnosis and fault-tolerant control method to fill the current technological gap, ensure that the steer-by-wire system meets ASIL D functional safety requirements, and guarantee its stable and reliable operation. Summary of the Invention

[0005] In view of this, to fill the gaps and deficiencies in existing technologies, this invention proposes a fault diagnosis and fault-tolerant control method and system for a dual-winding permanent magnet synchronous motor. This invention aims to solve the safety and steering angle tracking performance problems of steer-by-wire systems. By constructing mathematical models of three common faults in dual-winding permanent magnet synchronous motors (including single-phase open circuit, winding grounding, and inter-turn short circuit), typical fault characteristics are determined. For the characteristics of different fault types, a rapid diagnosis method based on current residuals and an intelligent diagnosis method based on time-frequency domain fusion features and CNN are designed to achieve accurate fault identification and information feedback. Based on the fault diagnosis results, corresponding fault-tolerant control strategies are triggered. Combined with a non-singular fast terminal sliding mode control algorithm, effective fault isolation and fault-tolerant control are performed according to the fault type after a fault occurs, improving the safety and stability of the dual-winding permanent magnet synchronous motor.

[0006] This invention proposes a fault diagnosis and fault-tolerant control method and system for a dual-winding permanent magnet synchronous motor, including the following:

[0007] This invention proposes a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor, characterized by the following:

[0008] Step S1: For the target motor, which is a dual-winding permanent magnet synchronous motor, design diagnostic methods for open circuit faults, ground faults, and inter-turn short circuit faults, perform fault diagnosis on the target motor, and output motor fault status flags.

[0009] Step S2: Based on the motor fault status flag obtained in Step S1, distinguish the corresponding motor fault status; if the target motor is fault-free, perform angle tracking control on the fault-free target motor; if the target motor is in a fault state, trigger the corresponding fault-tolerant control scheme according to different fault states.

[0010] Step S3: Based on the control scheme determined in step S2, apply the corresponding constraints and solve for the control current. If it is in a healthy state, the control current is determined only by the motor model of the target motor and the angle tracking controller, without any additional constraints. If it is in a fault state, apply additional constraints to the control current based on the fault type on the basis of the healthy state to obtain the fault-tolerant control current, so that the dual-winding permanent magnet synchronous motor can operate stably in both healthy and fault states.

[0011] Further, step S1 includes the following:

[0012] Step S11: First, acquire the three-phase current signals of the two windings of the target motor using a current sensor, and perform signal processing to eliminate sensor temperature drift and ensure the validity of fault characteristics. The sampling period is kept consistent with the control period of the target motor to ensure signal timing synchronization. To address the inherent zero-point drift problem of the sensor, zero-point calibration is performed in the static state before the target motor starts, and the drift compensation value is calculated by continuously acquiring raw current signals for multiple cycles. The zero-point compensation function is:

[0013] ;

[0014] In the formula, Let be the sampled current of phase x at time k. Here, M represents the phase current after zero-value compensation, and M represents the number of sampling periods for zero-value compensation. This refers to the motor torque;

[0015] Step S12: To suppress high-frequency noise introduced by PWM modulation and electromagnetic interference, and to avoid residual calculation distortion caused by signal phase lag, synchronous low-pass filtering is performed on the calibrated sampling current and the reference current output by the current loop controller. The filtering formulas are as follows:

[0016] Sampling current filtering:

[0017] ;

[0018] In the formula The sampled current after filtering. These are the filter coefficients;

[0019] Reference current filtering:

[0020] ;

[0021] In the formula This is the filtered reference current;

[0022] Step S13: Based on the preprocessed current signal, calculate the normalized phase current residual; in the process of calculating the normalized phase current residual, the residual is converted into a dimensionless relative value. The formula for calculating the normalized residual is:

[0023] ;

[0024] In the formula Let be the residual current of phase x at time k.

[0025] Furthermore, step S1 also includes the following:

[0026] Step S14: Based on the triple judgment logic of amplitude threshold, duration period and dynamic shielding, design the open circuit fault judgment function of the target motor; wherein the fault judgment function adopts the triple judgment logic of amplitude threshold, duration period and dynamic shielding.

[0027] The first step is to determine whether the single-cycle residual is greater than a set threshold by using an indicator function.

[0028] The second step involves using a summation function to determine whether the duration of the residual error exceeding the limit is greater than a set period. When the duration of the residual error exceeding the limit is greater than the set period, the residual error exceeding the limit is a persistent fault.

[0029] The third judgment is designed with a dynamic shielding function, which uses the motor speed change rate as the identification basis to avoid misjudgment caused by the motor's current loop response speed being limited, resulting in large transient errors and increased residuals.

[0030] Open circuit fault determination function The design is as follows:

[0031] ;

[0032] Where N is the number of periods for continuous determination. For indicator functions, their function expression is:

[0033] ;

[0034] In the formula The set threshold for determining open-circuit faults;

[0035] design This is a dynamic masking function, based on the rate of change of rotational speed. The function expression for identifying whether the system is in a wide range of transient conditions is as follows:

[0036] ;

[0037] In the formula The set threshold for the rate of change of rotational speed.

[0038] Furthermore, step S1 also includes the following:

[0039] Step S15: Design the ground fault determination function for the target motor under ground fault conditions based on the zero-sequence current component; the calculation formula for the zero-sequence current residual is as follows:

[0040] ;

[0041] Designed ground fault determination function as follows:

[0042] ;

[0043] Among them, the grounding indication function The expression is as follows:

[0044] ;

[0045] In the formula The set threshold for determining ground faults.

[0046] Further, step S1 includes the following:

[0047] Step S16: Design a diagnostic scheme based on time-frequency domain feature fusion and lightweight CNN to achieve accurate identification of inter-turn short-circuit fault levels, including the following:

[0048] Frequency domain features are extracted from the current signal using fast Fourier transform, and then fused with statistically obtained time domain features to form a 10-dimensional feature vector, which is then input into the CNN model.

[0049] The designed CNN network structure is as follows: The input layer receives a multi-dimensional feature matrix; the hidden layer contains three cascaded convolutional blocks, each following an architecture from a convolutional layer to a batch normalization layer to a ReLU activation layer to a max pooling layer, compressing the feature dimension while retaining key information; the output of the convolutional blocks is flattened and then fed into a fully connected layer. The first fully connected layer has several neurons, followed by a cascaded Dropout layer to suppress overfitting; the second fully connected layer has four neurons, corresponding to four fault levels; the output layer uses a Softmax function to convert the neuron outputs into probabilities for each level, and takes the maximum probability as the diagnostic result. .

[0050] Further, step S2 includes the following:

[0051] Step S21: Determine the fault flag bit obtained in step S1. The target motor is in a healthy state when the following formula is satisfied:

[0052] ;

[0053] If the target motor is in a healthy state, it can be directly subjected to angle tracking control without applying additional constraints to the current.

[0054] Step S22: Determine the fault flag obtained in step S1. The target motor is in a single-phase fault state when the following formula is satisfied:

[0055] ;

[0056] When the expression in step S22 is satisfied, the target motor is in a phase open circuit or inter-turn short circuit fault. The corresponding fault-tolerant control scheme is: to disconnect the main circuit of the phase bridge arm by the circuit breaker and disconnect the power supply of the phase to isolate the phase winding and prevent the fault phase from further expanding its impact; at the same time, connect the midpoint of the two capacitors on the DC side to the stator neutral point to use this as a voltage reference point to make up for the lack of voltage constraint after the phase loss.

[0057] Step S23: Determine the fault flag obtained in step S1. The motor is in a winding fault state when the following formula is satisfied:

[0058] ;

[0059] When the expression in step S23 is satisfied, the corresponding fault-tolerant control scheme is as follows: the winding containing the faulty phase is isolated as a whole, and then the main circuit of the winding is cut off by the circuit breaker to disconnect the power supply to the winding to isolate the fault.

[0060] Further, step S3 includes the following:

[0061] Step S31: For the target motor in a healthy state, a non-singular fast terminal sliding mode controller was designed to perform angular tracking control, including the following:

[0062] The error is defined as the difference between the reference angle and the actual angle, and the formula is:

[0063] ;

[0064] According to the kinematics of an electric motor, the second-order dynamic equation of the error is:

[0065] ;

[0066] In the formula This refers to the moment of inertia of the motor. This represents the number of pole pairs of the motor. For permanent magnet flux linkage; The damping coefficient; This represents the motor load torque.

[0067] The designed sliding surface function is:

[0068] ;

[0069] In the formula Design parameters for the sliding surface; It is a power factor and satisfies the constraints. ;

[0070] The designed sliding mode reaching law is:

[0071] ;

[0072] In the formula, For exponential convergence rate; To switch the gain; It is a smoothing factor;

[0073] Based on the above formula, the sliding mode control law at this time can be derived as follows:

[0074] ;

[0075] ;

[0076] Step S32: Calculate the q-axis reference current of each of the two windings of the target motor in a healthy state; in a healthy state, the q-axis currents of the two windings of the motor are equal and both are... ,Right now:

[0077] ;

[0078] Step S33: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S32, realize the rotation angle tracking control;

[0079] Step S34: When the target motor has a single-phase fault, calculate the fault-tolerant control phase current under the single-phase fault state according to the additional constraints under the single-phase fault in step S2; the expression for the motor dq-axis current under the healthy state is known to be:

[0080] ;

[0081] ;

[0082] ;

[0083] ;

[0084] In the formula These are the d-axis currents of the first and second sets of windings, respectively. These are the q-axis currents of the first and second windings, respectively. The electric angle of the motor.

[0085] Furthermore, step S3 also includes the following:

[0086] Step S35: Use two-phase four-switch modulation instead of the normal three-phase six-switch modulation strategy for the winding where the faulty phase is located;

[0087] Step S36: If it is a winding fault, calculate the fault-tolerant control current under winding isolation according to the additional constraints under single-phase fault in step S2; since one set of windings is isolated at this time, and the neutral points of the two sets of windings are independent of each other, the original dual-winding permanent magnet synchronous motor degenerates into an ordinary three-phase permanent magnet synchronous motor, and its constraints are:

[0088] ;

[0089] Right now = The current in the healthy winding doubles, bearing the entire torque output;

[0090] Step S37: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S36, realize the rotation angle tracking control.

[0091] According to a second aspect of the present invention, a fault diagnosis and fault-tolerant control system for a dual-winding permanent magnet synchronous motor includes an electronic device, wherein the electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor as described in any one of the present invention.

[0092] According to a third aspect of the present invention, a fault diagnosis and fault-tolerant control system for a dual-winding permanent magnet synchronous motor includes a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor as described in any one of the present invention.

[0093] The present invention has the following advantages:

[0094] This invention enables steer-by-wire systems to diagnose three typical and frequently occurring faults. After fault identification, by isolating the fault and actively reconstructing the topology, combined with appropriate fault-tolerant control algorithms, the system achieves continuous and reliable fault-tolerant operation. This invention significantly improves the functional safety level of steer-by-wire systems while providing superior steering angle tracking control accuracy. Attached Figure Description

[0095] Figure 1 This is a schematic diagram of the steps of the present invention.

[0096] Figure 2 This is a schematic diagram of the process of the present invention.

[0097] Figure 3 This is a diagram of the CNN structure in this invention. Detailed Implementation

[0098] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0099] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0100] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments of the present invention; as used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise; furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0101] like Figures 1 to 3 As shown, this invention proposes a fault diagnosis and fault-tolerant control method and system for a dual-winding permanent magnet synchronous motor, characterized by including the following:

[0102] This invention proposes a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor, characterized by the following:

[0103] Step S1: Design diagnostic methods for open circuit faults, ground faults, and inter-turn short circuit faults, perform fault diagnosis on a dual-winding permanent magnet synchronous motor, and output motor fault status flags;

[0104] Step S2: Based on the motor fault status flags obtained in Step S1 (including fault type and specific fault phase sequence), distinguish the corresponding motor fault status; if the motor is fault-free, perform angle tracking control as a normal motor. If the motor is in a fault state, trigger the corresponding fault-tolerant control scheme according to different fault states;

[0105] Step S3: Based on the control scheme determined in Step S2, apply the corresponding constraints and solve for the control current. In a healthy state, the control current is determined solely by the motor model and the angle tracking controller, without additional constraints. In a fault state, additional constraints are applied to the control current based on the fault type, in addition to the healthy state constraints, to obtain a fault-tolerant control current, ensuring stable operation of the dual-winding permanent magnet synchronous motor in both healthy and fault states.

[0106] In one embodiment of the present invention, step S1 specifically includes:

[0107] Step S11: First, acquire the three-phase current signals of the two windings of the dual-winding permanent magnet synchronous motor using a current sensor, and perform signal processing to eliminate sensor temperature drift and ensure the validity of fault characteristics; the sampling period is kept consistent with the motor control period to ensure signal timing synchronization; to address the inherent zero-point drift problem of the sensor, perform zero-point calibration in the static state before motor startup, and continuously acquire raw current signals for multiple cycles to calculate the drift compensation value; the zero-value compensation function is:

[0108] ;

[0109] In the formula, Let be the sampled current of phase x at time k. Here, M represents the phase current after zero-value compensation, and M represents the number of sampling periods for zero-value compensation. This refers to the motor torque;

[0110] Step S12: To suppress high-frequency noise introduced by PWM modulation and electromagnetic interference, and to avoid residual calculation distortion caused by signal phase lag, synchronous low-pass filtering is performed on the calibrated sampling current and the reference current output by the current loop controller. The filtering formulas are as follows:

[0111] Sampling current filtering:

[0112] ;

[0113] In the formula The sampled current after filtering. These are the filter coefficients;

[0114] Reference current filtering:

[0115] ;

[0116] In the formula This is the filtered reference current;

[0117] Step S13: Based on the preprocessed current signal, calculate the normalized phase current residual. Considering that the residual amplitude will change with the motor load and operating conditions during motor operation, it is necessary to normalize the residual to facilitate the setting of the diagnostic threshold, converting the residual into a dimensionless relative value. The formula for calculating the normalized residual is:

[0118] ;

[0119] In the formula Let be the current residual of phase x at time k;

[0120] Step S14: Based on the triple judgment logic of amplitude threshold, duration period and dynamic shielding, design an open circuit fault judgment function; wherein the fault judgment function adopts the triple judgment logic of amplitude threshold, duration period and dynamic shielding.

[0121] The first layer of judgment uses an indicator function to determine whether the single-cycle residual exceeds a set threshold. The second layer uses a summation function to determine whether the duration of the residual exceeding the limit exceeds a set period, thus confirming that the residual exceeding the limit is a persistent fault rather than an occasional signal distortion. The third layer of judgment uses a dynamic shielding function, based on the motor speed change rate, to avoid misjudgments caused by the limited current loop response speed leading to larger transient errors and increased residuals under conditions such as motor start-up, acceleration / deceleration, and sudden load changes. This three-layer judgment logic establishes a complete fault diagnosis chain from three dimensions: residual amplitude, duration period, and system operational stability.

[0122] For open-circuit faults, the core fault characteristic is that the current in one phase abnormally returns to zero, which will generate a huge phase current residual error. Approaching 1. Once The system immediately identifies the faulty phase and triggers the corresponding fault-tolerant control strategy. Open-circuit fault determination function. The design is as follows:

[0123] ;

[0124] Where N is the number of periods for continuous determination. For indicator functions, their function expression is:

[0125] ;

[0126] In the formula The threshold for determining open-circuit faults is set.

[0127] This is a dynamic masking function, based on the rate of change of rotational speed. The function expression for identifying whether the system is in a wide range of transient conditions is as follows:

[0128] ;

[0129] In the formula The set threshold for the rate of change of rotational speed.

[0130] Step S15: Based on the triple judgment logic of "amplitude threshold, duration period, and dynamic shielding", design an open-circuit fault judgment function. For ground faults, the core fault characteristic is that the three-phase current balance is disrupted, and the originally zero-sequence current component exhibits abnormal fluctuations. The formula for calculating the zero-sequence current residual is as follows:

[0131] ;

[0132] Designed ground fault determination function as follows:

[0133] ;

[0134] Among them, the grounding indication function The expression is as follows:

[0135] ;

[0136] In the formula The set threshold for determining ground faults.

[0137] Step S16: To address the issue of weak early characteristics of inter-turn short-circuit faults being easily masked by load fluctuations, a diagnostic scheme of "time-frequency domain feature fusion + lightweight CNN" is designed to achieve accurate identification of the inter-turn short-circuit fault level. Frequency domain features are extracted from the current signal using Fast Fourier Transform, and fused with statistically obtained time domain features to form a 10-dimensional feature vector, which is then input into the CNN model.

[0138] The designed lightweight CNN network structure is as follows: The input layer receives a 10×1×1 feature matrix; the hidden layer contains three cascaded convolutional blocks, each following the architecture of "convolutional layer → batch normalization layer → ReLU activation layer → max pooling layer". The first and second convolutional blocks use [3,1]-sized convolutional kernels with 16 and 32 filters respectively, while the third convolutional block uses [4,1]-sized convolutional kernels to expand the receptive field and has 64 filters; the pooling layers all use [2,1] max pooling kernels to retain key information while compressing the feature dimension; the output of the convolutional blocks is flattened and then fed into fully connected layers. The first fully connected layer has 64 neurons, followed by a cascaded dropout layer to suppress overfitting; the second fully connected layer has 4 neurons, corresponding to 4 fault levels; the output layer uses the Softmax function to convert the neuron outputs into probabilities for each level, and takes the maximum probability as the diagnostic result. .

[0139] In one embodiment of the present invention, step S2 specifically includes:

[0140] Step S21: Determine the fault flag obtained in step S1. If the following formula is satisfied, the motor is in a healthy state:

[0141] ;

[0142] As described in step S2, if the motor is in a healthy state, it can be directly subjected to angle tracking control without applying additional constraints to the current.

[0143] Step S22: Determine the fault flag obtained in step S1. If the following formula is satisfied, it indicates that the motor is in a single-phase fault state:

[0144] ;

[0145] This state indicates that the motor is experiencing an open circuit or inter-turn short circuit fault in one phase. In this case, the main circuit of the bridge arm of that phase is disconnected by the circuit breaker, and the power supply to that phase is cut off to isolate the fault in the winding of that phase and prevent the faulty phase from further expanding its impact. At the same time, the midpoint of the two capacitors on the DC side is connected to the stator neutral point to serve as a voltage reference point to compensate for the lack of voltage constraint after the phase loss.

[0146] Step S23: Determine the fault flag obtained in step S1. If the following formula is satisfied, it indicates that the motor is in a winding fault state:

[0147] ;

[0148] In this state, since the ground fault has a significant impact on the three-phase current, the winding containing the faulty phase must be isolated as a whole. This can be achieved by cutting off the main circuit of the winding and disconnecting the power supply to the winding through a circuit breaker.

[0149] In one embodiment of the present invention, step S3 specifically includes:

[0150] Step S31: According to step S2, in the healthy state, direct angle tracking control is performed. This invention designs a non-singular fast terminal sliding mode controller for angle tracking control, defining the error as the difference between the reference angle and the actual angle, as shown in the formula:

[0151] ;

[0152] According to the kinematics of an electric motor, the second-order dynamic equation of the error is:

[0153] ;

[0154] In the formula This refers to the moment of inertia of the motor. This represents the number of pole pairs of the motor. For permanent magnet flux linkage; The damping coefficient; This represents the motor load torque.

[0155] The designed sliding surface function is:

[0156] ;

[0157] In the formula Design parameters for the sliding surface; It is a power factor and satisfies the constraints.

[0158] The designed sliding mode reaching law is:

[0159] ;

[0160] In the formula, For exponential convergence rate; To switch the gain; This is a smoothing factor.

[0161] Based on the above formula, the sliding mode control law at this time can be derived as follows:

[0162] ;

[0163] ;

[0164] Step S32: Calculate the q-axis reference current for each of the two windings under healthy conditions. In the ideal operating condition of a healthy motor, each winding bears half of the torque output, meaning the q-axis current of both windings is equal and both are... ,Right now:

[0165] ;

[0166] Step S33: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S32, realize the rotation angle tracking control.

[0167] Step S34: If it is a single-phase fault, calculate the fault-tolerant control phase current under the single-phase fault state according to the additional constraints under the single-phase fault in step S2. The expression for the motor dq-axis current under healthy state is known to be:

[0168] ;

[0169] ;

[0170] ;

[0171] ;

[0172] In the formula These are the d-axis currents of the first and second sets of windings, respectively. These are the q-axis currents of the first and second windings, respectively. The electric angle of the motor.

[0173] Since the motor output torque is proportional to the q-axis current, to ensure that the motor torque does not fluctuate significantly after a fault, the q-axis current must not drop sharply due to the current returning to zero after a phase is isolated. Taking a phase A fault as an example, its isolation essentially adds the following constraint to the system:

[0174] ;

[0175] With the above constraints added, each of the two windings bears half of the torque output, maintaining... If the solution remains unchanged, we can obtain:

[0176] ;

[0177] Step S35: Use two-phase four-switch modulation instead of the normal three-phase six-switch modulation strategy for the winding where the faulty phase is located.

[0178] Step S36: If it is a winding fault, calculate the fault-tolerant control current under winding isolation based on the additional constraints under single-phase fault in Step S2. Since one set of windings is isolated at this time, and the neutral points of the two sets of windings are independent of each other, the original dual-winding permanent magnet synchronous motor degenerates into a common three-phase permanent magnet synchronous motor, with the following constraints:

[0179] ;

[0180] Right now = The current in the healthy winding doubles, bearing the entire torque output.

[0181] Step S37: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S36, realize the rotation angle tracking control.

[0182] The above are preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention shall fall within the protection scope of the present invention.

Claims

1. A fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor, characterized in that, Includes the following: Step S1: For the target motor, which is a dual-winding permanent magnet synchronous motor, design diagnostic methods for open circuit faults, ground faults, and inter-turn short circuit faults, perform fault diagnosis on the target motor, and output motor fault status flags. Step S2: Distinguish the corresponding motor fault status based on the motor fault status flag obtained in step S1; If the target motor is fault-free, then perform angle tracking control on the fault-free target motor; If the target motor is in a fault state, the corresponding fault-tolerant control scheme will be triggered according to the different fault states. Step S3: Based on the control scheme determined in step S2, apply the corresponding constraints and solve for the control current; If the motor is in a healthy state, the control current is determined solely by the target motor's motor model and the angle tracking controller, without any additional constraints. If the fault condition is met, additional constraints are applied to the control current based on the fault type, in addition to the control current under healthy conditions, to obtain a fault-tolerant control current, so that the dual-winding permanent magnet synchronous motor can operate stably under both healthy and fault conditions.

2. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 1, characterized in that, Step S1 includes the following: Step S11: First, the three-phase current signals of the two windings of the target motor are acquired through the current sensor, and the signals are processed to eliminate sensor temperature drift and ensure the effectiveness of fault characteristics. The sampling period is kept consistent with the control period of the target motor to ensure signal timing synchronization. To address the inherent zero-point drift problem of the sensor, zero-point calibration is performed in the static state before the target motor starts, and the drift compensation value is calculated by continuously acquiring raw current signals for multiple cycles. The zero-point compensation function is: ； In the formula, Let be the sampled current of phase x at time k. The phase current after zero-value compensation is M, the number of sampling cycles for zero-value compensation is Te, and the motor torque is Te. Step S12: To suppress high-frequency noise introduced by PWM modulation and electromagnetic interference, and to avoid residual calculation distortion caused by signal phase lag, synchronous low-pass filtering is performed on the calibrated sampling current and the reference current output by the current loop controller. The filtering formulas are as follows: Sampling current filtering: ； In the formula The sampled current after filtering. These are the filter coefficients; Reference current filtering: ； In the formula This is the filtered reference current; Step S13: Calculate the normalized phase current residual based on the preprocessed current signal; In calculating the normalized phase current residual, the residual is converted into a dimensionless relative value. The formula for calculating the normalized residual is as follows: ； In the formula Let be the residual current of phase x at time k.

3. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 2, characterized in that, Step S1 also includes the following: Step S14: Based on the triple judgment logic of amplitude threshold, duration period and dynamic shielding, design the open circuit fault judgment function of the target motor; wherein the fault judgment function adopts the triple judgment logic of amplitude threshold, duration period and dynamic shielding. The first step is to determine whether the single-cycle residual is greater than a set threshold by using an indicator function. The second step involves using a summation function to determine whether the duration of the residual error exceeding the limit is greater than a set period. When the duration of the residual error exceeding the limit is greater than the set period, the residual error exceeding the limit is a persistent fault. The third judgment is designed with a dynamic shielding function, which uses the motor speed change rate as the identification basis to avoid misjudgment caused by the motor's current loop response speed being limited, resulting in large transient errors and increased residuals. Open circuit fault determination function The design is as follows: ； Where N is the number of periods for continuous determination. For indicator functions, their function expression is: ； In the formula The set threshold for determining open-circuit faults; design This is a dynamic masking function, based on the rate of change of rotational speed. The function expression for identifying whether the system is in a wide range of transient conditions is as follows: ； In the formula The set threshold for the rate of change of rotational speed.

4. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 3, characterized in that, Step S1 also includes the following: Step S15: Design the ground fault determination function for the target motor under ground fault conditions based on the zero-sequence current component; the calculation formula for the zero-sequence current residual is as follows: ； Designed ground fault determination function as follows: ； Among them, the grounding indication function The expression is as follows: ； In the formula The set threshold for determining ground faults.

5. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 4, characterized in that, Step S1 includes the following: Step S16: Design a diagnostic scheme based on time-frequency domain feature fusion and lightweight CNN to achieve accurate identification of inter-turn short-circuit fault levels, including the following: Frequency domain features are extracted from the current signal using fast Fourier transform, and then fused with statistically obtained time domain features to form a multi-dimensional feature matrix, which is then input into a CNN model. The designed CNN network structure is as follows: The input layer receives a multi-dimensional feature matrix; the hidden layer contains three cascaded convolutional blocks, each following an architecture from a convolutional layer to a batch normalization layer to a ReLU activation layer to a max pooling layer, compressing the feature dimension while retaining key information; the output of the convolutional blocks is flattened and then fed into a fully connected layer. The first fully connected layer has several neurons, followed by a cascaded Dropout layer to suppress overfitting; the second fully connected layer has four neurons, corresponding to four fault levels; the output layer uses a Softmax function to convert the neuron outputs into probabilities for each level, and takes the maximum probability as the diagnostic result. .

6. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 5, characterized in that, Step S2 includes the following: Step S21: Determine the fault flag bit obtained in step S1. The target motor is in a healthy state when the following formula is satisfied: ； If the target motor is in a healthy state, it can be directly subjected to angle tracking control without applying additional constraints to the current. Step S22: Determine the fault flag obtained in step S1. The target motor is in a single-phase fault state when the following formula is satisfied: ； When the expression in step S22 is satisfied, the target motor is in a phase open circuit or inter-turn short circuit fault. The corresponding fault-tolerant control scheme is: to disconnect the main circuit of the phase bridge arm by the circuit breaker and disconnect the power supply of the phase to isolate the phase winding and prevent the fault phase from further expanding its impact; at the same time, connect the midpoint of the two capacitors on the DC side to the stator neutral point to use this as a voltage reference point to make up for the lack of voltage constraint after the phase loss. Step S23: Determine the fault flag obtained in step S1. The motor is in a winding fault state when the following formula is satisfied: ; When the expression in step S23 is satisfied, the corresponding fault-tolerant control scheme is as follows: the winding containing the faulty phase is isolated as a whole, and then the main circuit of the winding is cut off by the circuit breaker to disconnect the power supply to the winding to isolate the fault.

7. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 6, characterized in that, Step S3 includes the following: Step S31: For the target motor in a healthy state, a non-singular fast terminal sliding mode controller was designed to perform angular tracking control, including the following: The error is defined as the difference between the reference angle and the actual angle, and the formula is: ； According to the kinematics of an electric motor, the second-order dynamic equation of the error is: ； In the formula This refers to the moment of inertia of the motor. This represents the number of pole pairs of the motor. For permanent magnet flux linkage; The damping coefficient; This represents the motor load torque. The designed sliding surface function is: ； In the formula Design parameters for the sliding surface; It is a power factor and satisfies the constraints. ; The designed sliding mode reaching law is: ； In the formula, For exponential convergence rate; To switch the gain; It is a smoothing factor; Based on the above formula, the sliding mode control law at this time can be derived as follows: ； ； Step S32: Calculate the q-axis reference current of each of the two windings of the target motor in a healthy state; in a healthy state, the q-axis currents of the two windings of the motor are equal and both are... ,Right now: ； Step S33: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S32, realize the rotation angle tracking control; Step S34: When the target motor has a single-phase fault, calculate the fault-tolerant control phase current under the single-phase fault state according to the additional constraints under the single-phase fault in step S2; the expression for the motor dq-axis current under the healthy state is known to be: ； ； ； ； In the formula These are the d-axis currents of the first and second sets of windings, respectively. These are the q-axis currents of the first and second windings, respectively. The electric angle of the motor.

8. The fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor according to claim 7, characterized in that, Step S3 also includes the following: Step S35: Use two-phase four-switch modulation instead of the normal three-phase six-switch modulation strategy for the winding where the faulty phase is located; Step S36: If it is a winding fault, calculate the fault-tolerant control current under winding isolation according to the additional constraints under single-phase fault in step S2; since one set of windings is isolated at this time, and the neutral points of the two sets of windings are independent of each other, the original dual-winding permanent magnet synchronous motor degenerates into an ordinary three-phase permanent magnet synchronous motor, and its constraints are: ； Right now = The current in the healthy winding doubles, bearing the entire torque output; Step S37: Based on the q-axis reference current input current loop controllers of the two sets of windings obtained in step S36, realize the rotation angle tracking control.

9. A fault diagnosis and fault-tolerant control system for a dual-winding permanent magnet synchronous motor, comprising an electronic device, wherein the electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor as described in any one of claims 1 to 8.

10. A fault diagnosis and fault-tolerant control system for a dual-winding permanent magnet synchronous motor, comprising a computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a fault diagnosis and fault-tolerant control method for a dual-winding permanent magnet synchronous motor as described in any one of claims 1 to 8.

Images