A two-stage current reconfiguration fault-tolerant control method for dual three-phase permanent magnet synchronous motors
By employing a two-stage current reconfiguration fault-tolerant control method for dual three-phase permanent magnet synchronous motors, the d-axis double frequency trajectory and subspace current mapping are optimized, solving the problems of torque ripple and current distribution under a single neutral point open circuit fault, and achieving optimized distribution of healthy phase current and suppression of torque ripple.
Patent Information
- Application Number
- CN202611443665.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-17
- Publication Date
- 2026-11-17
AI Technical Summary
After an open-circuit fault occurs in a single-neutral-point dual three-phase permanent magnet synchronous motor, existing technologies cannot effectively utilize the zero-sequence current adjustment degree of freedom, resulting in torque pulsation and suboptimal current distribution. Furthermore, traditional compensation methods fail to fully consider the impact of changes in the d-axis current trajectory and mapping coefficient on the current distribution.
A two-stage current reconfiguration fault-tolerant control method for dual three-phase permanent magnet synchronous motors is proposed. By setting the d-axis second harmonic trajectory adjustment component and the subspace current mapping coefficient for joint optimization, a first-level fault-tolerant current reference is generated. Then, based on the remaining pulsating torque, the characteristic harmonic compensation current of the q-axis is calculated to perform a second-level current reconfiguration, so as to improve the current distribution of healthy phases and suppress torque pulsation.
It achieves optimized current distribution in healthy phases, reduced stator copper losses, improved torque output capability, and suppressed torque pulsation under the conditions of zero fault phase current and single neutral point current closure, and is applicable to existing motor drive systems.
Smart Images

Figure CN123122407A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of multiphase permanent magnet synchronous motor drive and fault-tolerant control technology, specifically relating to a two-stage current reconfiguration fault-tolerant control method applicable to open-circuit faults in single-neutral-point dual three-phase permanent magnet synchronous motors. Background Technology
[0002] The dual three-phase permanent magnet synchronous motor consists of two sets of three-phase stator windings with a spatial phase difference of 30° electrical angle. It features high power density, low torque ripple, multiple degrees of control freedom, and strong fault tolerance, making it suitable for high-reliability drive applications such as ship electric propulsion, aerospace, and new energy vehicles. When an open-circuit fault occurs in part of the stator winding, the remaining healthy phases can continue to establish a rotating magnetic field, and fault-tolerant control can maintain motor operation. Depending on the connection relationship of the neutral points of the two sets of three-phase windings, the dual three-phase permanent magnet synchronous motor can adopt a dual-neutral-point isolated structure or a single-neutral-point connection structure. The single-neutral-point structure provides a common zero-sequence current path for both sets of windings, increasing the degree of freedom in current regulation after a fault, but also creating a more complex coupling relationship between the fundamental, harmonic, and zero-sequence subspaces. When an open-circuit fault occurs in any phase, the fault phase current is restricted to zero, and the currents of the remaining phases must also satisfy the instantaneous current closure condition of the common neutral point. Therefore, the orthogonal current subspaces that operate normally cannot be configured completely independently.
[0003] According to vector space decoupling theory, the six-phase currents of a dual three-phase permanent magnet synchronous motor can be mapped to the fundamental subspace, harmonic subspace, and zero-sequence subspace. Traditional fault-tolerant control maintains torque output after a fault by configuring reference currents for each subspace to satisfy the conditions of zero fault phase current and neutral point current closure. Among these, fault-tolerant methods based on sinusoidal phase currents impose constraints on the current waveform, allowing for further optimization of current distribution after a fault.
[0004] Existing research has focused on dual-neutral-point isolation structures. This involves introducing a component that varies with twice the rotor electrical angle period into the d-axis current reference of the synchronous rotating coordinate system, and optimizing its amplitude, phase, and harmonic subspace current mapping coefficients to improve healthy phase current distribution. In this structure, two sets of three-phase windings each satisfy their respective neutral-point current closure conditions, and both zero-sequence subspace currents are zero.
[0005] For a structure with two sets of three-phase windings connected to the same neutral point, the two zero-sequence subspace currents are allowed to be non-zero and satisfy the relationship of equal magnitude and opposite direction. The current constraints and available current distribution degrees of freedom after a fault are different from those of a double-neutral-point isolated structure. Therefore, under the constraint of a single-neutral-point open-circuit fault, how to utilize the zero-sequence current adjustment degree of freedom to determine the fault current distribution adapted to the d-axis second harmonic injection, and how to match the subsequent torque compensation with the changed current distribution, are technical problems that need to be solved.
[0006] In practical dual-phase three-phase permanent magnet synchronous motors, the permanent magnet flux linkage and back electromotive force typically contain third, fifth, and seventh spatial harmonics. The third spatial harmonic primarily maps to the zero-sequence subspace, while the fifth and seventh spatial harmonics primarily map to the harmonic subspace. Under a single neutral-point open-circuit fault condition, the zero-sequence and harmonic subspace currents generated to meet fault constraints interact with the aforementioned higher-order back electromotive forces, producing pulsating components in the output torque at two, four, and six times the electrical angular frequency. Existing torque pulsation suppression methods typically inject characteristic harmonic compensation components into the q-axis current reference, generating a compensation torque with a phase opposite to the original torque pulsation. While this method can reduce torque fluctuations after a fault, the compensation current is usually calculated based on a predetermined fault-tolerant current distribution, failing to fully consider the changes in the amplitude, phase, and residual torque pulsation composition of the currents in each subspace after the d-axis current trajectory and mapping coefficients have changed.
[0007] When d-axis current trajectory optimization and q-axis torque compensation are designed separately and directly superimposed, they will interact: d-axis injection and mapping coefficient optimization will change the harmonic and zero-sequence subspace currents, causing the q-axis compensation current obtained from the original current distribution to no longer accurately correspond to the actual residual pulsating torque; after coordinate transformation, the q-axis compensation current will re-participate in the current distribution of the fundamental, harmonic, and zero-sequence subspaces, further altering the effective value, peak value, and stator copper loss of each healthy phase current. Therefore, simply superimposing the two current components is insufficient to simultaneously maintain the current optimization result and torque compensation effect. It is necessary to design a correlation between post-fault current distribution optimization and residual torque compensation, ensuring that the fault phase current is zero and the single neutral point current is closed, while also considering healthy phase current distribution, stator copper loss, torque output capability, and torque pulsation performance. Summary of the Invention
[0008] Objective: To address the current distribution and torque ripple issues following an open-circuit fault in a single-neutral-point dual-three-phase permanent magnet synchronous motor, this invention proposes a two-stage current reconfiguration fault-tolerant control method for dual-three-phase permanent magnet synchronous motors. Under the conditions of zero fault phase current and single-neutral-point current closure constraint, the method sets a d-axis second harmonic frequency trajectory adjustment component using the q-axis fundamental current reference as the amplitude benchmark. Its amplitude coefficient, phase angle, and current mapping coefficients in the harmonic subspace and zero-sequence subspace are all incorporated into the optimization, forming the first-level fault-tolerant current references for the minimum copper loss mode and the maximum torque mode, respectively. The corresponding torque compensation current is calculated based on the first-level current reference, and the second-level current reconfiguration is performed through the optimized current mapping relationship to improve the current distribution of the healthy phase and suppress torque ripple.
[0009] To achieve the above objectives, this invention provides a two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor with two sets of three-phase windings connected to the same neutral point. This method generates a first-layer fault-tolerant current reference through joint optimization of the d-axis trajectory adjustment component and the subspace current mapping coefficient. It then calculates the q-axis characteristic harmonic compensation current based on the remaining pulsating torque after the first-layer reconfiguration, and finally substitutes the compensated q-axis current back into the optimized current mapping relationship to generate a second-layer fault-tolerant current reference. Specifically, the method includes the following steps: Step 1: Obtain the motor rotor electrical angle and stator current. Use vector space decoupling transformation to map the six-phase natural coordinate system variables to the fundamental subspace, harmonic subspace and zero-sequence subspace. Based on the fact that the fault phase current is zero and the single neutral point current closure condition, establish the current mapping relationship between the reference currents of the harmonic subspace and zero-sequence subspace and the reference current of the fundamental subspace. Among them, the two zero-sequence subspace reference currents satisfy the relationship of equal magnitude and opposite direction. Step 2: In the synchronous rotating coordinate system, using the q-axis fundamental current reference that generates a given average electromagnetic torque as the amplitude reference, a trajectory adjustment component that varies with twice the rotor electrical angle period is set in the d-axis current reference. The trajectory adjustment component is obtained by multiplying the q-axis fundamental current reference, the amplitude coefficient, and a second-harmonic sine function containing the phase angle. The amplitude coefficient, the phase angle, and the mapping coefficients used to configure the harmonic subspace and zero-sequence subspace reference currents in the current mapping relationship are used together as variables to be optimized. Under the constraints of zero fault phase current, equal magnitude and opposite direction of the two zero-sequence subspace reference currents, average electromagnetic torque maintained at a given value, and healthy phase current not exceeding the inverter's allowable current, the first layer of fault-tolerant current reference is obtained by jointly solving the variables to be optimized with the minimum stator total copper loss or the minimum effective value of the maximum current of the healthy phase as the fault-tolerant optimization objective. Step 3: Based on the first-layer fault-tolerant current reference, calculate the remaining pulsating torque components generated by the coupling of the third, fifth, and seventh back electromotive forces with the first-layer fault-tolerant current reference.
[0010] Step 4: Using the antiphase quantity of the sum of the remaining pulsating torque components as the compensation target, solve for the q-axis characteristic harmonic compensation current.
[0011] Step 5: Keeping the d-axis current reference and optimized current mapping coefficients determined in Step 2 unchanged, the q-axis characteristic harmonic compensation current is superimposed on the q-axis fundamental current reference before compensation, and the superimposed q-axis current reference and the d-axis current reference are subjected to Park inverse transformation to obtain the compensated fundamental subspace current reference; the compensated fundamental subspace current reference is substituted back into the current mapping relationship obtained in Step 2, and the harmonic subspace and zero-sequence subspace reference currents are recalculated so that the q-axis characteristic harmonic compensation current participates in the redistribution of subspace currents and healthy phase currents after the fault, and a second-layer fault-tolerant current reference dependent on the first-layer fault-tolerant current reference is obtained; Step 6: Under the condition that the current of each healthy phase corresponding to the second layer fault-tolerant current reference does not exceed the inverter's allowable current, perform closed-loop tracking of the second layer fault-tolerant current reference and generate the switching signal of the six-phase inverter according to the voltage command of each subspace.
[0012] Furthermore, in step 1, iron loss, magnetic circuit saturation, and changes in motor parameters with temperature and operating conditions are ignored, and the stator winding resistance, inductance, and permanent magnet flux linkage are assumed to remain constant.
[0013] Based on the principle of vector space decoupling, a Clarke transformation matrix suitable for the winding structure is constructed to map the six-phase natural coordinate system variables to mutually orthogonal α-β fundamental wave subspace, xy harmonic subspace, and o1-o2 zero-sequence subspace; combined with the Park transformation, the α-β fundamental wave subspace variables are projected onto the synchronously rotating dq coordinate system, thereby realizing the decoupled representation of each subspace variable.
[0014]
[0015]
[0016] In the formula, For rotor electrical angle, It is the transformation matrix from the original coordinate system to the stationary coordinate system. It is the transformation matrix from a stationary coordinate system to a rotating coordinate system.
[0017] Further, in step 1, a basic motor model is first established that only considers the fundamental flux linkage of the permanent magnets to describe the relationship between the fundamental current and torque; the additional electromagnetic torques corresponding to the third, fifth, and seventh permanent magnet spatial flux linkages are calculated separately in step 3. The basic model is as follows: Magnetic flux linkage equation:
[0018] In the formula, , These are the magnetic flux linkages along the d and q axes, respectively. , These are the magnetic flux linkages along the x and y axes, respectively. , These are the magnetic flux linkages along axes o1 and o2, respectively. and These are the d-axis and q-axis inductances, respectively. This represents the amplitude of the fundamental spatial flux linkage of the permanent magnet. The equivalent leakage inductance is given by the xy harmonic subspace and the o1-o2 zero-order subspace. and These are the d-axis and q-axis currents, respectively. and The currents along the x-axis and y-axis are respectively. , These are the currents of the o1 and o2 axes, respectively; Voltage equation:
[0019]
[0020] In the formula, and These are the voltages along the d-axis and q-axis, respectively. and The voltages along the x and y axes are respectively. and These are the voltages for the O1 and O2 axes, respectively. This refers to the stator winding resistance. Electric angular velocity; Electromagnetic torque equation:
[0021] In the formula, This is the electromagnetic torque corresponding to the fundamental magnetic flux and the magnetoresistive effect; It is the extreme logarithm; The aforementioned dual three-phase permanent magnet synchronous motor is a surface-mounted permanent magnet synchronous motor, and it is considered that... Therefore, neglecting the reluctance torque, the electromagnetic torque equation simplifies to:
[0022] Equations of motion:
[0023] And satisfy:
[0024] In the formula, It is the moment of inertia; This refers to the rotor's mechanical angular velocity; Electromagnetic torque; This is the load torque; is the viscous damping coefficient.
[0025] Furthermore, taking an open-circuit fault in phase F as the preset fault state, the fault phase current satisfies:
[0026] Since the two sets of three-phase windings are connected to the same neutral point, according to the single neutral point current closure condition, the two zero-sequence shaft currents in the first layer of fault-tolerant current reference satisfy:
[0027] In the formula, For phase F current, and These are the reference currents for the first zero-sequence axis and the second zero-sequence axis, respectively.
[0028] The dual three-phase permanent magnet synchronous motor includes a first set of three-phase windings and a second set of three-phase windings with a spatial phase difference of 30° electrical angle, and the first set of three-phase windings and the second set of three-phase windings are connected to the same neutral point. The vector space decoupling transformation maps the six-phase natural coordinate system variables to mutually orthogonal α-β fundamental subspaces, xy harmonic subspaces, and o1-o2 zero-sequence subspaces, and projects the α-β fundamental subspace variables onto the synchronously rotating dq coordinate system through the Park transformation; the current mapping relationship is expressed as:
[0029] In the formula, and These are the α-axis and β-axis reference currents in the first-level fault-tolerant current reference; , , and These are the x-axis, y-axis, o1-axis, and o2-axis reference currents in the first layer of fault-tolerant current reference; is the current mapping coefficient; the superscript "(0)" indicates a variable based on the first-level fault-tolerant current reference and without q-axis characteristic harmonic compensation. The superscript "*" indicates the reference value when used for current or voltage variables, and the optimized value when used for optimization variables or mapping coefficients.
[0030] According to the vector space decoupling transformation used in this embodiment, the F-phase current satisfies:
[0031] Substitute the above mapping relationship:
[0032] To ensure that the F-phase current remains zero throughout the entire electrical cycle, the coefficient should satisfy:
[0033] Under the constraints of the above equations, the current mapping coefficients, together with the amplitude coefficients and phase angles of the d-axis second harmonic trajectory adjustment component, participate in the optimization.
[0034] The aforementioned current mapping relationship is established based on a structure where two sets of three-phase windings are connected to the same neutral point. Unlike the constraint in a dual-neutral-point isolation structure where both zero-sequence subspace currents are zero, this structure allows the two zero-sequence subspace currents to participate in post-fault current distribution under the condition that they are equal in magnitude and opposite in direction. Therefore, when setting the d-axis second harmonic trajectory adjustment component with the q-axis fundamental current reference as the amplitude base, the harmonic subspace and zero-sequence subspace current mapping coefficients, along with the amplitude coefficients and phase angles of the trajectory adjustment component, are jointly incorporated into the optimization and solved jointly under the constraints of zero fault phase current and single neutral-point current closure.
[0035] Furthermore, a d-axis trajectory adjustment component that varies with twice the rotor electrical angle period is set in the d-axis current reference of the synchronous rotating coordinate system. The d-axis trajectory adjustment component is expressed as:
[0036] In the formula, To generate a q-axis fundamental current reference for a given average electromagnetic torque, The amplitude coefficient of the d-axis trajectory adjustment component. Adjust the phase angle of the component for the d-axis trajectory. The rotor electrical angle; In this process, the second-layer current reconstruction keeps the d-axis current reference unchanged and does not recalculate the d-axis trajectory adjustment component with the compensated q-axis current reference. Projecting the d-axis trajectory adjustment component and the q-axis fundamental current reference onto the α-β fundamental subspace using the inverse Park transform, we obtain:
[0037] Substituting the d-axis trajectory adjustment component into the above formula, we further obtain:
[0038] The amplitude coefficient, phase angle, and current mapping coefficient of the trajectory adjustment component are used as the variables to be optimized:
[0039] By performing an inverse vector space decoupling transformation on the first-level fault-tolerant current reference, we obtain the K-phase instantaneous current reference: In the formula, , Let α be the electrical angle between the axis of the K-phase winding and the α-axis.
[0040] The effective value of the K-phase current is:
[0041] When minimizing stator copper loss is the fault-tolerant optimization objective, the variables to be optimized are solved according to the following formula:
[0042] Given an average electromagnetic torque, the current distribution optimization objective for the maximum torque mode is to minimize the effective value of the maximum current in the healthy phase.
[0043] In the formula, ~ This refers to the effective value of the phase current of phases A through F.
[0044] Within the operating range where the current reference varies proportionally with the torque, the effective value of the phase current is the main limiting factor, and the voltage and instantaneous current peak limits are not reached, the smaller the maximum effective value of the healthy phase current under the given average electromagnetic torque, the larger the available effective value margin of the current. Therefore, the target is used to evaluate and optimize the torque output capability under the effective value constraint. In the formula, The vector of variables to be optimized; and These are the optimal variable vectors corresponding to the minimum copper loss mode and the maximum torque mode, respectively; the subscripts "ML" and "MT" represent the minimum copper loss mode and the maximum torque mode, respectively. This refers to the stator winding resistance. All of the above optimizations satisfy the condition that the fault phase current is zero. The constraints include the average electromagnetic torque being equal to the given torque and the healthy phase current not exceeding the inverter's allowable current. and These are the reference currents for the first zero-sequence axis and the second zero-sequence axis, respectively.
[0045] Furthermore, the dual three-phase permanent magnet synchronous motor is a surface-mounted permanent magnet synchronous motor, and its d-axis inductance... With q-axis inductance satisfy In step 3, the fundamental, third, fifth, and seventh spatial flux linkage amplitudes of the permanent magnet are pre-acquired and stored in the controller as motor characteristic parameters, which can be obtained through finite element calculation or no-load back EMF identification. Using the first-layer fault-tolerant current reference as the basis for calculating the remaining pulsating torque, the instantaneous electromagnetic powers generated by the first, third, fifth, and seventh back EMFs in their respective subspaces are as follows:
[0046] When the rotor mechanical angular velocity At that time, the total electromagnetic torque corresponding to the first layer fault-tolerant current reference is expressed as:
[0047] In the formula, , , and These are the instantaneous electromagnetic powers generated by the fundamental, third, fifth, and seventh back electromotive forces and the first-layer fault-tolerant current reference effect, respectively. , , , , , , and These are the components of each back electromotive force in the corresponding subspace; The total electromagnetic torque corresponding to the first layer fault-tolerant current reference; ω is the rotor mechanical angular velocity.
[0048] The back electromotive force and the space flux linkage of the permanent magnet satisfy the following:
[0049] In the formula, It is the back electromotive force. This represents the number of pole pairs of the motor. The rotor's mechanical angular velocity, For permanent magnet space flux linkage, The rotor electrical angle is given.
[0050] The third-order magnetic flux linkage is projected onto the zero-order subspace o1-o2, the fifth-order magnetic flux linkage is projected onto the xy harmonic subspace in reverse order rotation, and the seventh-order magnetic flux linkage is projected onto the xy harmonic subspace in forward order rotation. The corresponding back electromotive force components are:
[0051] In the formula, This represents the amplitude of the fundamental spatial flux linkage of the permanent magnet. , , These represent the amplitudes of the third, fifth, and seventh spatial flux linkages of the permanent magnet, respectively. The third-order spatial flux linkage is projected onto the zero-order subspace o1-o2, the fifth-order spatial flux linkage is projected onto the xy harmonic subspace in reverse order rotation, and the seventh-order spatial flux linkage is projected onto the xy harmonic subspace in forward order rotation; the residual pulsating torque components caused by the third, fifth, and seventh-order back electromotive forces are expressed as follows:
[0052] In the formula, , and These are the residual pulsating torque components generated by the coupling of the third, fifth, and seventh back electromotive forces with the first-layer fault-tolerant current reference, respectively; the subscripts 3, 5, and 7 are used to indicate the back electromotive force harmonic sources of the corresponding residual pulsating torque components, and do not indicate the order of torque pulsation. This represents the number of pole pairs of the motor. , and These represent the amplitudes of the third, fifth, and seventh spatial flux linkages of the permanent magnet, respectively.
[0053] The aforementioned residual pulsating torque components are all calculated using the first-layer fault-tolerant current reference, thus already including the influence of the d-axis trajectory adjustment component and the optimized current mapping coefficient on the harmonic subspace, zero-sequence subspace current, and residual torque pulsation. Therefore, the q-axis characteristic harmonic compensation current obtained is not determined based on the traditional fault-tolerant current distribution, but rather corresponds to the actual residual pulsating torque after the first-layer current reconstruction.
[0054] Furthermore, in step 4, the q-axis current reference is decomposed into a q-axis fundamental current reference that generates a given average electromagnetic torque. and q-axis characteristic harmonic compensation current used to cancel the residual pulsating torque .
[0055] Let the compensation torque generated by the q-axis characteristic harmonic compensation current have the same amplitude and opposite phase as the sum of the third, fifth, and seventh residual pulsating torques:
[0056] Therefore, we can conclude that:
[0057] The compensated q-axis current reference is:
[0058] Right now:
[0059] Furthermore, in step 5, the compensated q-axis current reference and the d-axis trajectory adjustment current determined by the first layer reconstruction are subjected to a Park inverse transformation:
[0060] Will and Substituting back into the optimized current mapping relationship, we obtain the second-level fault-tolerant current reference:
[0061] In the formula, and These are the α-axis and β-axis reference currents in the second-layer fault-tolerant current reference; , , and These are the x, y, o1, and o2 axis reference currents in the second-layer fault-tolerant current reference; The optimized current mapping coefficients obtained in step 2 are used; the second layer fault-tolerant current reference is recalculated from the q-axis characteristic harmonic compensation current through the optimized current mapping relationship. The second layer current reconstruction uses the current mapping coefficients obtained from the first layer optimization and keeps the d-axis current reference determined in the first layer unchanged.
[0062] The second-layer current reconstruction uses the current mapping coefficients obtained in the first layer and keeps the d-axis current reference determined in the first layer unchanged. The compensated q-axis current reference is then substituted back using the Park inverse transform and the aforementioned current mapping relationship to generate the final current references for each subspace. The phase current changes caused by the second-layer compensation are included in the performance evaluation of the final current RMS value and stator copper losses.
[0063] Let the maximum ratio of the higher-order space flux linkage to the fundamental space flux linkage of the permanent magnet be:
[0064] In the formula, It is the maximum ratio of the higher-order spatial flux linkage of the permanent magnet to the fundamental spatial flux linkage, and is a dimensionless quantity.
[0065] when At that time, the q-axis characteristic harmonic compensation current is about The first-order quantity, whose subspace current increment caused by backsubstitution and subsequent coupling with higher-order magnetic flux results in an additional torque, is about The second-order quantity. Therefore, in the analytical calculation of the compensation current, the present invention ignores the second-order additional torque and still determines the q-axis characteristic harmonic compensation current based on the residual pulsating torque corresponding to the first-layer fault-tolerant current reference.
[0066] Depend on , , , , and Together they constitute the second layer of fault-tolerant current reference. Within the operating range set by this invention, the current of each healthy phase corresponding to the second layer of fault-tolerant current reference does not exceed the inverter's allowable current, and can be directly used for closed-loop tracking control.
[0067] The second-level reconstruction does not directly superimpose the compensation current at the phase current end, but instead substitutes the compensated q-axis current back into the fundamental subspace transformation relationship and the optimized current mapping relationship, so that the final reference current still satisfies the fault phase current being zero and the single neutral point current closing condition.
[0068] Further, in step 6, the second-layer fault-tolerant current reference is used as the current closed-loop setpoint for the fundamental subspace, harmonic subspace, and zero-sequence subspace, respectively. The actual six-phase stator current is obtained by vector space decoupling transformation. The corresponding subspace voltage command is generated according to the deviation between the reference current and the actual current of each subspace. The subspace voltage command is then processed by vector space decoupling inverse transformation and modulation to generate the switching signal of the six-phase inverter.
[0069] Specifically, in step 6, closed-loop tracking is performed on the second-layer fault-tolerant current reference. Vector space decoupling transformation and Park transformation are performed on the actual six-phase stator currents to obtain the actual current in each subspace; the actual current is compared with the corresponding reference current, and voltage commands are obtained through the current regulators in each subspace. , , , , and .
[0070] The instantaneous voltage command for phase K is:
[0071] In the formula, The instantaneous voltage commands for each phase are input into the modulation module to generate the switching signals for the six-phase inverter, enabling fault-tolerant operation of the dual three-phase permanent magnet synchronous motor under open-circuit fault conditions.
[0072] Furthermore, when an open-circuit fault occurs in any of phases A, B, C, D, E, or F, corresponding fault phase current constraints and current mapping relationships are established based on the spatial location of the faulty phase. The current mapping coefficients, d-axis trajectory adjustment component amplitude coefficients, and phase angles corresponding to this fault state are solved. The first-level fault-tolerant current reference is generated using corresponding optimization variables, and subsequent two-stage current reconstruction and closed-loop control are executed. The values of the optimization variables differ for different faults, but the first-level current reconstruction, residual pulsating torque calculation, q-axis characteristic harmonic compensation, second-level current reconstruction, and closed-loop control processes remain consistent.
[0073] Compared with existing open-circuit fault-tolerant control methods for dual three-phase permanent magnet synchronous motors, the present invention has the following advantages: 1) This invention addresses open-circuit faults in single-neutral-point dual three-phase permanent magnet synchronous motors. Utilizing the structural feature that allows zero-sequence current to participate in post-fault current distribution, it sets a d-axis second harmonic trajectory adjustment component using the q-axis fundamental current reference as the amplitude benchmark. The amplitude coefficient, phase angle, and current mapping coefficients of the harmonic subspace and zero-sequence subspace are all incorporated into the optimization. Under the constraints of zero fault phase current and single-neutral-point current closure, the invention jointly solves for the objectives of minimizing total stator copper loss and minimizing the effective value of the maximum current in the healthy phase. This achieves a healthy phase current distribution that matches the corresponding fault-tolerant objectives, providing an optimized current reference for reducing copper loss or improving torque output capability under current effective value constraints.
[0074] 2) Based on the first-layer fault-tolerant current reference, this invention determines the remaining pulsating torque components caused by the third, fifth, and seventh back electromotive forces, so that the obtained q-axis characteristic harmonic compensation current corresponds to the actual remaining pulsating torque after the first-layer current reconstruction, thereby improving the pertinence of torque pulsation compensation.
[0075] 3) The present invention substitutes the compensated q-axis current reference back to the current mapping relationship obtained by the first layer optimization, so that the q-axis characteristic harmonic compensation current participates in the redistribution of the fundamental subspace, harmonic subspace, zero-sequence subspace and healthy phase current. While suppressing torque pulsation, the final reference current continues to satisfy the fault phase current being zero and the single neutral point current closing condition.
[0076] 4) This invention unifies fault current distribution optimization and residual torque ripple compensation into a two-stage current reconfiguration framework, without changing the hardware structure of the motor and inverter, and can use the original vector control and current closed-loop structure, making it easy to implement in existing dual three-phase motor drive systems. Attached Figure Description
[0077] Figure 1 This is a block diagram illustrating the two-stage current reconfiguration principle of the present invention. Figure 2 This is a block diagram of the control strategy for the two-stage current reconfiguration fault-tolerant control method for dual three-phase permanent magnet synchronous motors of the present invention. Figure 3 The simulation results of phase current under the minimum copper loss mode in phase F open-circuit fault are presented. Figure 3 (a) represents the traditional fault-tolerant method. Figure 3 (b) is the method of the present invention; Figure 4 The simulation comparison results show the maximum healthy phase current RMS value under the minimum copper loss mode during an open-circuit fault in phase F. Figure 4 (a) represents the traditional fault-tolerant method. Figure 4 (b) is the method of the present invention; Figure 5 The simulation results of electromagnetic torque are compared for the minimum copper loss mode under phase F open-circuit fault. Figure 5 (a) represents the traditional fault-tolerant method. Figure 5 (b) is the method of the present invention; Figure 6 The simulation results of phase current under the maximum torque mode of phase F open-circuit fault are presented for comparison. Figure 6 (a) represents the traditional fault-tolerant method. Figure 6 (b) is the method of the present invention; Figure 7 The simulation comparison results show the maximum healthy phase current RMS value under the maximum torque mode during an open-circuit fault in phase F. Figure 7 (a) represents the traditional fault-tolerant method. Figure 7 (b) is the method of the present invention; Figure 8 The electromagnetic torque simulation comparison results are shown for the maximum torque mode under the F-phase open-circuit fault. Figure 8 (a) represents the traditional fault-tolerant method. Figure 8 (b) is the method of the present invention; Detailed Implementation
[0078] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the invention will be further described below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments described are only for explaining the invention and are not intended to limit the scope of protection of the invention.
[0079] This embodiment uses a dual three-phase permanent magnet synchronous motor with two sets of three-phase windings connected to the same neutral point as the controlled object. The two sets of three-phase windings are spatially 30° electrical degrees apart and are powered by a six-phase inverter. Taking an open-circuit fault in phase F as an example, according to... Figure 1 The two-stage current reconstruction principle shown and Figure 2 The control strategy shown implements fault-tolerant control. This embodiment uses a surface-mounted permanent magnet synchronous motor and meets the following requirements. Therefore, electromagnetic torque does not include reluctance torque component.
[0080] Step 1: Establish the mathematical model of the motor and the mapping relationship of open-circuit fault current. Collect six-phase stator current and rotor electrical angle and rotor mechanical angular velocity By employing vector space decoupling transformation, the variables of the six-phase natural coordinate system are mapped to mutually orthogonal α-β fundamental wave subspace, xy harmonic wave subspace, and o1-o2 zero-order subspace:
[0081]
[0082] Projecting the α-β fundamental wavelet space currents onto the synchronously rotating dq coordinate system using the Park transform:
[0083] In the formula, For rotor electrical angle, It is the transformation matrix from the original coordinate system to the stationary coordinate system.
[0084] Ignoring iron losses, magnetic circuit saturation, and parameter variations, a basic motor model is first established, considering only the fundamental flux linkage of the permanent magnets. The effects of the spatial flux linkages of the third, fifth, and seventh permanent magnets are separately considered in step 3. The flux linkage equation of the basic model is as follows:
[0085] The voltage equation is:
[0086]
[0087] In the formula, For stator winding resistance, and These are the d-axis and q-axis inductances, respectively. The equivalent leakage inductance for the xy harmonic subspace and the o1-o2 zero-order subspace is given. This represents the amplitude of the fundamental spatial flux linkage of the permanent magnet. ω is the rotor's electric angular velocity.
[0088] Electromagnetic torque equation:
[0089] In the formula, This is the electromagnetic torque corresponding to the fundamental magnetic flux and the magnetoresistive effect; It is an extreme logarithm.
[0090] This embodiment uses a surface-mounted permanent magnet synchronous motor, and it is believed that... Therefore, neglecting the reluctance torque, the electromagnetic torque equation simplifies to:
[0091] Equations of motion:
[0092] And satisfy:
[0093] In the formula, It is the moment of inertia; This refers to the rotor's mechanical angular velocity; Electromagnetic torque; This is the load torque; is the viscous damping coefficient.
[0094] After phase F is opened, the following conditions are met:
[0095] Since the two sets of three-phase windings are connected to the same neutral point, according to the single neutral point current closure condition, the first layer of fault-tolerant current reference satisfies:
[0096] Substituting the fault constraints into the vector space decoupling inverse transformation relationship, the first layer of fault-tolerant current mapping relationship is established:
[0097] In the formula, each current variable is the corresponding subspace reference current in the first layer of fault-tolerant current reference; The current mapping coefficient to be optimized; the superscript "(0)" indicates a variable based on the first layer fault-tolerant current reference and without q-axis characteristic harmonic compensation; the superscript "*" indicates a reference value when used for current or voltage variables, and an optimized value when used for optimization variables or mapping coefficients.
[0098] According to the vector space decoupling transformation used in this embodiment, the F-phase current satisfies:
[0099] Substitute the above mapping relationship:
[0100] To ensure that the F-phase current remains zero throughout the entire electrical cycle, the coefficient should satisfy:
[0101] Under the constraints of the above equations, the current mapping coefficients, together with the amplitude coefficients and phase angles of the d-axis second harmonic trajectory adjustment component, participate in the optimization.
[0102] The aforementioned current mapping relationship is established based on a structure where two sets of three-phase windings are connected to the same neutral point. Unlike the constraint in a dual-neutral-point isolation structure where both zero-sequence subspace currents are zero, this structure allows the two zero-sequence subspace currents to participate in post-fault current distribution under the condition that they are equal in magnitude and opposite in direction. Therefore, when setting the d-axis second harmonic trajectory adjustment component with the q-axis fundamental current reference as the amplitude base, the harmonic subspace and zero-sequence subspace current mapping coefficients, along with the amplitude coefficients and phase angles of the trajectory adjustment component, are jointly incorporated into the optimization and solved jointly under the constraints of zero fault phase current and single neutral-point current closure.
[0103] Step 2: Generate the first layer of fault-tolerant current reference The speed regulator generates a q-axis fundamental current reference to maintain a given average electromagnetic torque based on the deviation between the given speed and the actual speed. .
[0104] Set the d-axis trajectory adjustment component in the d-axis current reference as it varies with twice the rotor electrical angle period:
[0105] In the formula, The amplitude coefficient of the trajectory adjustment component. The phase angle of the trajectory adjustment component.
[0106] in, This is the fundamental current reference before the q-axis torque compensation current is superimposed. The second-layer current reconstruction keeps the d-axis current reference unchanged and does not recalculate the d-axis trajectory adjustment component based on the compensated q-axis current reference.
[0107] Will and Projected onto the α-β fundamental wavelet space via the inverse Park transform:
[0108] Substituting the d-axis trajectory adjustment component into the above formula, we further obtain:
[0109] The amplitude coefficient, phase angle, and current mapping coefficient of the trajectory adjustment component are used as the variables to be optimized:
[0110] By performing an inverse vector space decoupling transformation on the first-level fault-tolerant current reference, we obtain the K-phase instantaneous current reference:
[0111] In the formula, , Let α be the electrical angle between the axis of the K-phase winding and the α-axis.
[0112] The effective value of the K-phase current is:
[0113] When minimizing stator copper loss is the fault-tolerant optimization objective, the variables to be optimized are solved according to the following formula:
[0114] Given an average electromagnetic torque, the current distribution optimization objective for the maximum torque mode is to minimize the effective value of the maximum current in the healthy phase.
[0115] In the formula, ~ This refers to the effective value of the phase current of phases A through F.
[0116] Within the operating range where the current reference varies proportionally with the torque, the effective value of the phase current is the main limiting factor, and the voltage and instantaneous current peak limits are not reached, the smaller the effective value of the maximum healthy phase current under a given average electromagnetic torque, the larger the available effective current margin. Therefore, the aforementioned objective is used to evaluate and optimize the torque output capability under effective value constraints.
[0117] All of the above optimizations satisfy the condition that the fault phase current is zero. The constraints include the average electromagnetic torque being equal to the given torque and the healthy phase current not exceeding the inverter's allowable current.
[0118] This embodiment employs a sequential quadratic programming method for offline numerical solution of the variables to be optimized. Under the constraints of zero fault phase current, single neutral point current closure, average electromagnetic torque equal to the given torque, and healthy phase current not exceeding the inverter's allowable current, the optimization parameters corresponding to the minimum copper loss mode and the maximum torque mode are solved respectively. The minimum copper loss mode aims to minimize the total stator copper loss, while the maximum torque mode aims to minimize the effective value of the maximum current in the healthy phase. The obtained optimization parameters serve as preset control parameters for the corresponding fault conditions, used to generate the first-level fault-tolerant current reference. The effective value of the phase current and the stator copper loss in the objective function are calculated based on the first-level fault-tolerant current reference, without considering the phase current changes caused by the second-level q-axis compensation. The minimum copper loss mode and the maximum torque mode represent the optimization objectives adopted for the first-level current distribution.
[0119] For six states of single-phase open-circuit faults occurring in phases A, B, C, D, E, and F, the amplitude coefficient, phase angle, harmonics, and zero-sequence subspace current mapping coefficient of the d-axis second harmonic trajectory adjustment component are incorporated into the optimization based on the zero current of the corresponding fault phase and the single neutral point current closure condition. The minimum copper loss mode optimization parameters, obtained with the goal of minimizing total stator copper loss, are shown in Table 1; the maximum torque mode optimization parameters, obtained with the goal of minimizing the effective value of the maximum current of the healthy phase under a given average electromagnetic torque, are shown in Table 2. Both tables use the unified phase sequence, coordinate transformation, and rotor electrical angle definitions presented in this paper. The listed parameters are used to generate the first layer of fault-tolerant current reference before the superposition of q-axis compensation current.
[0120] Table 1. Optimization parameters for the first-level current reconfiguration of the minimum copper loss mode under each single-phase open-circuit fault. Fault Kd φd k1 k2 k3 k4 k5 k6 Phase A 0.25 0 -0.67 0 0 0 -0.33 0 Phase B -0.25 -π / 3 -0.17 0.29 -0.29 0.5 0.17 -0.29 C phase -0.25 π / 3 -0.17 -0.29 0.29 0.5 0.17 0.29 D phase 0.25 -π / 3 0.5 0.29 -0.29 -0.17 0.29 0.17 E phase 0.25 π / 3 0.5 -0.29 0.29 -0.17 -0.29 0.17 Phase F -0.25 0 0 0 0 -0.67 0 -0.33 Table 2. Optimization parameters for the first-level current reconfiguration of the maximum torque mode under various single-phase open-circuit faults. Fault Kd φd k1 k2 k3 k4 k5 k6 Phase A 0.454 -0.325 -0.69 0 -0.28 -0.12 -0.31 0 Phase B 0.454 1.77 0.04 0.14 -0.42 0.61 0.15 -0.27 C phase 0.454 -2.419 -0.2 -0.56 0.28 0.37 0.16 0.27 D phase 0.454 -0.723 0.37 0.28 -0.56 -0.2 0.27 0.16 E phase 0.454 1.372 0.61 -0.42 0.14 0.04 -0.27 0.15 Phase F 0.454 -2.817 -0.12 -0.28 0 -0.69 0 -0.31 Tables 1 and 2 present the first-level current reconstruction parameters for each single-phase open-circuit fault under two optimization objectives. During operation, the corresponding parameters are called based on the faulty phase and the selected fault-tolerant mode to generate the d-axis current reference and harmonic and zero-sequence subspace current references. Subsequent torque compensation is calculated based on the first-level current distribution corresponding to the selected parameters. The second-level current reconstruction uses the same set of current mapping coefficients and keeps the d-axis current reference determined in the first level unchanged. The following uses an open-circuit fault in phase F as an example to illustrate the subsequent control process.
[0121] Step 3: Calculate the remaining pulsating torque after the first layer reconstruction This embodiment uses a surface-mount dual three-phase permanent magnet synchronous motor, whose d-axis inductance and q-axis inductance satisfy... The reluctance torque term is zero. The following is a detailed derivation of the residual pulsating torque and the q-axis characteristic harmonic compensation current based on this implementation condition.
[0122] In this embodiment, the fundamental, third, fifth, and seventh spatial flux linkage amplitudes of the permanent magnet are pre-identified motor characteristic parameters stored in the controller. Specifically, the back electromotive force (EMF) of the motor is acquired under no-load operation with a non-zero rotor mechanical angular velocity, and Fourier decomposition is performed on the back EMF to obtain the amplitudes of the fundamental, third, fifth, and seventh back EMFs. For the h-th harmonic, its spatial flux linkage amplitude is determined according to the following formula:
[0123] In the formula, Let h be the amplitude of the back electromotive force. This represents the number of pole pairs of the motor. To identify the rotor mechanical angular velocity of the motor during the identification process. The identified... , , and The values are pre-stored in the controller and used to calculate the remaining pulsating torque after the first-level current reconstruction during operation. The amplitudes of each spatial flux linkage can also be pre-obtained through finite element analysis of the motor.
[0124] Using the first-layer fault-tolerant current reference as the basis for calculation, the instantaneous electromagnetic powers generated by the first, third, fifth, and seventh back electromotive forces in the corresponding subspaces are as follows:
[0125] The back electromotive force and the space flux satisfy:
[0126] The third-order spatial flux is projected onto the zero-order subspace of o1-o2, the fifth-order spatial flux is projected onto the xy harmonic subspace in a reverse rotation manner, and the seventh-order spatial flux is projected onto the xy harmonic subspace in a forward rotation manner.
[0127]
[0128] In the formula, , and These represent the residual pulsating torque components generated by the coupling of the third, fifth, and seventh back electromotive forces with the first-layer fault-tolerant current reference, respectively. The subscripts 3, 5, and 7 indicate the source of the back electromotive force harmonics corresponding to the residual pulsating torque, rather than the order of the torque pulsation. This represents the amplitude of the fundamental spatial flux linkage of the permanent magnet. , , These represent the amplitudes of the third, fifth, and seventh spatial flux linkages of the permanent magnet, respectively.
[0129] The aforementioned residual pulsating torque components are all calculated using the first-layer fault-tolerant current reference, thus already including the influence of the d-axis trajectory adjustment component and the optimized current mapping coefficient on the harmonic subspace, zero-sequence subspace current, and residual torque pulsation. Therefore, the q-axis characteristic harmonic compensation current obtained is not determined based on the traditional fault-tolerant current distribution, but rather corresponds to the actual residual pulsating torque after the first-layer current reconstruction.
[0130] Step 4: Solve for the q-axis characteristic harmonic compensation current The q-axis current reference is decomposed into a q-axis fundamental current reference that generates a given average electromagnetic torque. and characteristic harmonic compensation current used to cancel out residual pulsating torque .
[0131] Let the sum of the electromagnetic torque generated by the compensation current and the residual pulsating torque satisfy:
[0132] Therefore, we can conclude that:
[0133] The compensated q-axis current reference is:
[0134] Right now:
[0135] Step 5: Generate the second-layer fault-tolerant current reference The d-axis current reference determined by the first layer reconstruction With the compensated q-axis current reference Perform the inverse Park transform together:
[0136] Will and Substitute back to the optimized current mapping relationship:
[0137] The second-layer current reconstruction uses the current mapping coefficients obtained in the first layer and keeps the d-axis current reference determined in the first layer unchanged. The compensated q-axis current reference is then substituted back using the Park inverse transform and the aforementioned current mapping relationship to generate the final current references for each subspace. The phase current changes caused by the second-layer compensation are included in the performance evaluation of the final current RMS value and stator copper losses.
[0138] Let the maximum ratio of the higher-order flux linkage to the fundamental flux linkage of the permanent magnet be:
[0139] In the formula, It is the maximum ratio of the higher-order spatial flux linkage of the permanent magnet to the fundamental spatial flux linkage, and is a dimensionless quantity.
[0140] In this embodiment, the ratio of the higher-order flux linkage to the fundamental flux linkage of the permanent magnet satisfies the following condition: Therefore, the additional torque generated after the q-axis compensation current is about The second-order small quantity is ignored in the calculation of compensation current.
[0141] This yields the second-level fault-tolerant current reference. Under the torque and operating conditions set in this embodiment, the currents of each healthy phase corresponding to the second-level fault-tolerant current reference do not exceed the inverter's allowable current; therefore, no additional limiting is required for the q-axis characteristic harmonic compensation current.
[0142] Step 6: Implement current closed-loop control according to Figure 1 The schematic diagram of the two-stage current reconstruction shown below and Figure 2 The overall control strategy block diagram shown will and As a closed-loop reference for the dq-axis current, , , and As a closed-loop reference for harmonics and zero-sequence subspace current.
[0143] Based on the adopted post-fault current closed-loop structure, the corresponding current deviation is adjusted to obtain the voltage commands for the dq axis, harmonics, and zero-sequence subspace. The dq axis voltage command is then subjected to an inverse Park transform to obtain the fundamental subspace voltage command. and Then, it undergoes a vector space decoupling inverse transformation together with the harmonic and zero-sequence subspace voltage commands. The K-phase instantaneous voltage command is expressed as:
[0144] The six-phase voltage command is input into the modulation module to generate the switching control signal for the six-phase inverter, enabling fault-tolerant operation under open-circuit fault conditions.
[0145] This embodiment uses an open-circuit fault in phase F as an example. When an open-circuit fault occurs in any other phase, a corresponding fault constraint and current mapping relationship are established based on the spatial location of the faulty phase, and the optimized variable corresponding to the fault state is adopted. Different faulty phases correspond only to different values of the optimized variable; the current reconstruction and closed-loop control processes of the other two stages are the same as those for an open-circuit fault in phase F. Therefore, this invention is applicable to open-circuit faults in any phase of a dual three-phase permanent magnet synchronous motor.
[0146] Simulation results explanation: Figures 3-8 Simulation comparison results of the traditional fault-tolerant method and the two-stage current reconstruction fault-tolerant control method of this invention under an open-circuit fault in phase F are presented. In each figure, (a) represents the traditional fault-tolerant method, and (b) represents the method of this invention. Figures 3-5 The simulation results are as follows: phase current, RMS value of maximum healthy phase current, and electromagnetic torque under minimum copper loss mode; Figures 6-8 The figures show the simulation results of the phase current, the effective value of the maximum healthy phase current, and the electromagnetic torque under the maximum torque mode.
[0147] In contrast, traditional fault-tolerant methods employ conventional current distribution relationships corresponding to the minimum copper loss mode and the maximum torque mode, respectively, with the d-axis current reference set to zero and no q-axis torque compensation current configured. The method of this invention uses the optimized d-axis second harmonic trajectory adjustment parameters and current mapping coefficients obtained under the corresponding modes, and calculates the q-axis compensation current based on the first-layer current reference to complete the second-layer current reconstruction. The performance of the two methods is compared under the same given torque and other identical simulation conditions.
[0148] All simulations were conducted under the same given torque, speed, motor parameters, and current limits, employing consistent steady-state statistical intervals and sampling methods. Current distribution performance was evaluated using the effective value of phase current and stator copper losses, while torque fluctuation was evaluated using the relative standard deviation of electromagnetic torque. The relative standard deviation is the ratio of the torque standard deviation to the absolute value of the average torque, expressed as a percentage, where the standard deviation is normalized using the number of sampling points N. Under the same given torque, the torque output capability under effective value constraints is relatively improved based on the ratio of the maximum healthy effective value of phase current between the traditional method and the method of this invention.
[0149] Table 3. Simulation performance comparison between traditional fault-tolerant control methods and the two-stage current reconfiguration fault-tolerant control method of this invention.
[0150] As shown in Table 3, under the minimum copper loss mode, the stator copper loss of the method of the present invention is reduced from 43.64 W to 41.47 W, and the effective value of the maximum healthy phase current is reduced from 4.59 A to 4.02 A. Based on the effective value of the maximum healthy phase current, the torque output capability is relatively improved by approximately 14.18%, and the relative standard deviation of the electromagnetic torque is reduced from 6.589% to 1.082%. Under the maximum torque mode, the stator copper loss is reduced from 56.55 W to 43.68 W, and the effective value of the maximum healthy phase current is reduced from 3.53 A to 3.31 A. Based on the effective value of the maximum healthy phase current, the torque output capability is relatively improved by approximately 6.65%, and the relative standard deviation of the electromagnetic torque is reduced from 8.038% to 1.123%. The results show that, under the same given torque conditions, compared with the traditional fault-tolerant method without d-axis double frequency injection and q-axis torque compensation, the two-stage current reconstruction method of this invention reduces stator copper loss, the effective value of the maximum healthy phase current, and the relative standard deviation of electromagnetic torque, and improves the relative torque output capability calculated based on the effective value of current, demonstrating the comprehensive effect of this method in current distribution and torque fluctuation control.
[0151] The above embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Equivalent substitutions or modifications made by those skilled in the art without departing from the concept of the present invention should fall within the scope of protection of the present invention.
Claims
1. A two-stage current reconstruction fault-tolerant control method for a dual three-phase permanent magnet synchronous motor, characterized in that, The method for using a dual three-phase permanent magnet synchronous motor with two sets of three-phase windings connected to the same neutral point includes the following steps: Step 1: Obtain the rotor electrical angle and stator current of the motor. Use vector space decoupling transformation to map the six-phase natural coordinate system variables to the fundamental subspace, harmonic subspace, and zero-sequence subspace. Based on the zero fault phase current and the single neutral point current closure condition, establish the current mapping relationship between the reference currents of the harmonic subspace and zero-sequence subspace and the reference current of the fundamental subspace. Among them, the two zero-sequence subspace reference currents satisfy the relationship of equal magnitude and opposite direction. Step 2: In the synchronous rotating coordinate system, using the q-axis fundamental current reference that generates a given average electromagnetic torque as the amplitude reference, a trajectory adjustment component that varies with twice the rotor electrical angle period is set in the d-axis current reference. The trajectory adjustment component is obtained by multiplying the q-axis fundamental current reference, the amplitude coefficient, and a second-harmonic sine function containing the phase angle. The amplitude coefficient, the phase angle, and the mapping coefficients used to configure the harmonic subspace and zero-sequence subspace reference currents in the current mapping relationship are used together as variables to be optimized. Under the constraints of zero fault phase current, equal magnitude and opposite direction of the two zero-sequence subspace reference currents, average electromagnetic torque maintained at a given value, and healthy phase current not exceeding the inverter's allowable current, the first layer of fault-tolerant current reference is obtained by jointly solving the variables to be optimized with the minimum stator total copper loss or the minimum effective value of the maximum current of the healthy phase as the fault-tolerant optimization objective. Step 3: Based on the first-layer fault-tolerant current reference, calculate the remaining pulsating torque components generated by the third, fifth, and seventh back electromotive forces coupled with the first-layer fault-tolerant current reference, respectively. Step 4: Using the antiphase quantity of the sum of all remaining pulsating torque components as the compensation target, solve for the q-axis characteristic harmonic compensation current; Step 5: Keeping the d-axis current reference and optimized current mapping coefficients determined in Step 2 unchanged, the q-axis characteristic harmonic compensation current is superimposed on the q-axis fundamental current reference before compensation, and the superimposed q-axis current reference and the d-axis current reference are subjected to Park inverse transformation to obtain the compensated fundamental subspace current reference; the compensated fundamental subspace current reference is substituted back into the current mapping relationship obtained in Step 2, and the harmonic subspace and zero-sequence subspace reference currents are recalculated so that the q-axis characteristic harmonic compensation current participates in the redistribution of subspace currents and healthy phase currents after the fault, and a second-layer fault-tolerant current reference dependent on the first-layer fault-tolerant current reference is obtained; Step 6: Under the condition that the current of each healthy phase corresponding to the second layer fault-tolerant current reference does not exceed the inverter's allowable current, perform closed-loop tracking of the second layer fault-tolerant current reference and generate the switching signal of the six-phase inverter according to the voltage command of each subspace.
2. The two-stage current reconstruction fault-tolerant control method of dual three-phase permanent magnet synchronous motor according to claim 1, characterized in that, The dual three-phase permanent magnet synchronous motor includes a first set of three-phase windings and a second set of three-phase windings with a spatial phase difference of 30° electrical angle, and the first set of three-phase windings and the second set of three-phase windings are connected to the same neutral point. The vector space decoupling transformation maps the six-phase natural coordinate system variables to mutually orthogonal α-β fundamental subspaces, xy harmonic subspaces, and o1-o2 zero-sequence subspaces, and projects the α-β fundamental subspace variables onto the synchronously rotating dq coordinate system through the Park transformation; the current mapping relationship is expressed as: In the formula, and These are the α-axis and β-axis reference currents in the first-level fault-tolerant current reference; , , and These are the x-axis, y-axis, o1-axis, and o2-axis reference currents in the first layer of fault-tolerant current reference; For current mapping coefficients; the superscript "(0)" indicates a variable based on the first layer fault-tolerant current reference and without q-axis characteristic harmonic compensation. The superscript "*" indicates the reference value when used for current or voltage variables, and the optimized value when used for optimization variables or mapping coefficients.
3. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 2, characterized in that, In the d-axis current reference of the synchronous rotating coordinate system, a d-axis trajectory adjustment component that varies with twice the rotor electrical angle period is set. The d-axis trajectory adjustment component is expressed as: In the formula, To generate a q-axis fundamental current reference for a given average electromagnetic torque, The amplitude coefficient of the d-axis trajectory adjustment component. Adjust the phase angle of the component for the d-axis trajectory. The rotor electrical angle; In this process, the second-layer current reconstruction keeps the d-axis current reference unchanged and does not recalculate the d-axis trajectory adjustment component with the compensated q-axis current reference. Projecting the d-axis trajectory adjustment component and the q-axis fundamental current reference onto the α-β fundamental subspace using the inverse Park transform, we obtain: Substituting the d-axis trajectory adjustment component into the above formula, we further obtain: 。 4. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 3, characterized in that, The variable to be optimized in step 2 is represented as follows: By performing an inverse vector space decoupling transformation on the first-layer fault-tolerant current reference, the K-phase instantaneous current is obtained: In the formula, , The electrical angle between the axis of the K-phase winding and the α-axis; The effective value of the K-phase current is expressed as: When minimizing stator copper loss is the fault-tolerant optimization objective, the variables to be optimized are solved according to the following formula: Given the average electromagnetic torque, with the minimum effective value of the maximum current in the healthy phase as the equivalent optimization objective, solve for the variables to be optimized corresponding to the maximum torque mode: In the formula, ~ This refers to the effective value of the phase current of phases A through F; Within the operating range where the current reference varies proportionally with the torque, the effective value of the phase current is the main limiting factor, and the voltage and instantaneous current peak limits are not reached, the smaller the maximum effective value of the healthy phase current under the given average electromagnetic torque, the larger the available effective value margin of the current. Therefore, the target is used to evaluate and optimize the torque output capability under the effective value constraint. In the formula, The vector of variables to be optimized; and These are the optimal variable vectors corresponding to the minimum copper loss mode and the maximum torque mode, respectively; the subscripts "ML" and "MT" represent the minimum copper loss mode and the maximum torque mode, respectively. This refers to the stator winding resistance. All of the above optimizations satisfy the condition that the fault phase current is zero. The constraints include the average electromagnetic torque being equal to the given torque and the healthy phase current not exceeding the inverter's allowable current. and These are the reference currents for the first zero-sequence axis and the second zero-sequence axis, respectively.
5. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 4, characterized in that, The dual three-phase permanent magnet synchronous motor is a surface-mounted permanent magnet synchronous motor with a d-axis inductance. With q-axis inductance satisfy In step 3, the first-layer fault-tolerant current reference is used as the basis for calculating the remaining pulsating torque. The instantaneous electromagnetic power at the first, third, fifth, and seventh cycles is expressed as follows: When the rotor mechanical angular velocity At that time, the total electromagnetic torque corresponding to the first layer fault-tolerant current reference is expressed as: In the formula, , , and These are the instantaneous electromagnetic powers generated by the fundamental, third, fifth, and seventh back electromotive forces and the first-layer fault-tolerant current reference effect, respectively. , , , , , , and These are the components of each back electromotive force in the corresponding subspace; The total electromagnetic torque corresponding to the first layer fault-tolerant current reference; ω is the rotor mechanical angular velocity.
6. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 5, characterized in that, The third-order spatial flux linkage is projected onto the zero-order subspace o1-o2, the fifth-order spatial flux linkage is projected onto the xy harmonic subspace in reverse order rotation, and the seventh-order spatial flux linkage is projected onto the xy harmonic subspace in forward order rotation; the residual pulsating torque components caused by the third, fifth, and seventh-order back electromotive forces are expressed as follows: In the formula, , and These are the residual pulsating torque components generated by the coupling of the third, fifth, and seventh back electromotive forces with the first-layer fault-tolerant current reference, respectively; the subscripts 3, 5, and 7 are used to indicate the back electromotive force harmonic sources of the corresponding residual pulsating torque components, and do not indicate the order of torque pulsation. This represents the number of pole pairs of the motor. , and These represent the amplitudes of the third, fifth, and seventh spatial flux linkages of the permanent magnet, respectively.
7. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 6, characterized in that, Let the maximum ratio of the higher-order space flux linkage of the permanent magnet to the fundamental space flux linkage be: In the formula, It is the maximum ratio of the amplitude of the higher-order spatial flux linkage of the permanent magnet to the amplitude of the fundamental spatial flux linkage; when At that time, the q-axis characteristic harmonic compensation current is about The first-order quantity, whose subspace current increment caused by backsubstitution and the additional torque generated by its coupling with the higher-order space flux linkage are about The second-order quantity is ignored in the analytical calculation of the q-axis characteristic harmonic compensation current; In step 4, the q-axis current reference is decomposed into a q-axis fundamental current reference that generates a given average electromagnetic torque. and q-axis characteristic harmonic compensation current used to cancel out residual pulsating torque and establish: Therefore, we can conclude that: as well as: In the formula, This represents the amplitude of the fundamental spatial flux linkage of the permanent magnet. This is the final q-axis current reference after compensation; the subscript "inj" indicates the characteristic harmonic compensation component, and the subscript "ref" indicates the final reference value after the second layer reconstruction.
8. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 7, characterized in that, In step 5, the final q-axis current reference after q-axis characteristic harmonic compensation is substituted back into the fundamental subspace current transformation relationship to obtain: Then and Substituting back to the optimized current mapping relationship, we obtain the second-level fault-tolerant current reference: In the formula, and These are the α-axis and β-axis reference currents in the second-layer fault-tolerant current reference; , , and These are the x, y, o1, and o2 axis reference currents in the second-layer fault-tolerant current reference; The optimized current mapping coefficients obtained in step 2 are used; the second layer fault-tolerant current reference is recalculated from the q-axis characteristic harmonic compensation current through the optimized current mapping relationship. The second layer current reconstruction uses the current mapping coefficients obtained from the first layer optimization and keeps the d-axis current reference determined in the first layer unchanged.
9. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 8, characterized in that, In step 6, the second-layer fault-tolerant current reference is used as the current closed-loop setpoint for the fundamental subspace, harmonic subspace, and zero-sequence subspace, respectively. The actual six-phase stator current is obtained by vector space decoupling transformation. The corresponding subspace voltage command is generated according to the deviation between the reference current and the actual current of each subspace. The subspace voltage command is then processed by vector space decoupling inverse transformation and modulation to generate the switching signal of the six-phase inverter.
10. The two-stage current reconfiguration fault-tolerant control method for a dual three-phase permanent magnet synchronous motor according to claim 1, characterized in that, When an open-circuit fault occurs in any one of phases A, B, C, D, E, or F, the corresponding fault phase current constraints and current mapping relationships are established based on the spatial location of the fault phase. The current mapping coefficient, d-axis trajectory adjustment component amplitude coefficient, and phase angle corresponding to the fault state are solved. The first-level fault-tolerant current reference is generated using the corresponding optimization variables, and the subsequent two-stage current reconstruction and closed-loop control are executed.
