Fault-tolerant control system and method for multiphase permanent magnet synchronous motor

Abstract

The invention provides a fault-tolerant control system and method for a multiphase permanent magnet synchronous motor, relates to the technical field of motor control, and solves the limitation of the prior art on reliability, fault-tolerant performance and algorithm complexity. A fault confidence coefficient index is generated by fusing static and dynamic characteristics of the current, an active and robust fault diagnosis closed loop is constructed, and the reliability of a real working condition is improved; the method is not suitable for compromised degraded operation, but introduces a torque ripple suppression factor to optimize a current instruction, and combines a dynamic zero-sequence voltage component injection technology to maximize a bus voltage utilization rate, thereby ensuring high operation quality of all working conditions after a fault; the self-adaptive bandwidth harmonic suppression controller is adopted, the resonance bandwidth is adjusted on line through a lightweight bandwidth broadening factor, the calculation burden is remarkably reduced, and excellent engineering adaptability is shown by means of adjustable parameters.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, specifically to a fault-tolerant control system and method for a multiphase permanent magnet synchronous motor. Background Technology

[0002] Multiphase motors have more than three phases, such as five-phase and six-phase motors. By increasing the number of phases, redundant degrees of freedom are provided. When a phase fails, the remaining healthy phases can compensate for the function of the failed phase by redistributing the current or adjusting the magnetic field path.

[0003] The fault-tolerant control method for multiphase permanent magnet synchronous motors refers to the technical means of maintaining stable operation of the motor by adjusting the control algorithm when the motor fails, such as when the winding is open-circuited, short-circuited, or the sensor fails.

[0004] The existing technology, with publication number CN117526782B, entitled "A Fault-Tolerant Control Method for Multiphase Permanent Magnet Synchronous Motors Based on Voltage Limiting Analysis," includes: selecting vector combinations from the four quadrants of the fundamental plane under open-circuit fault conditions; determining the equivalence of the vector combinations with carrier-based pulse width modulation (CPWM); determining the maximum output voltage values ​​of the fundamental and harmonic planes based on the CPWM-equivalent vector combinations; calculating the voltage output limiting values ​​of the four quadrants of the fundamental plane considering harmonic suppression; comparing the magnitude of the reference voltage vector in the fundamental plane with the voltage output limiting values, and adjusting the reference voltage vector magnitudes of the fundamental and harmonic planes; and applying the adjusted reference voltage vectors of the fundamental and harmonic planes to the multiphase permanent magnet synchronous motor. This improves the operating efficiency in the normal modulation zone when the multiphase permanent magnet synchronous motor experiences an open-circuit fault and reduces the control error of the dq axes when the motor is operating in the over-modulation zone, ensuring the motor's control performance.

[0005] However, in practical applications, the above invention has the following technical drawbacks:

[0006] 1. In terms of reliability, the solution has limitations due to its passive and idealistic nature: This technology is a passive fault-tolerance strategy, which assumes that the fault has been accurately identified, but it does not include active fault diagnosis and rapid switching mechanisms, thus failing to form a complete reliability closed loop. In addition, this method relies heavily on accurate mathematical models for voltage limiting calculations, and is quite sensitive to changes in motor parameters due to actual operating conditions such as temperature rise and magnetic saturation, which can easily lead to control inaccuracies, thereby affecting the stability and true reliability of the system.

[0007] 2. Regarding performance after fault tolerance, its core is a compromise strategy of "downgrading to ensure operation": In order to ensure the basic power output of the motor under high load, this method will actively reduce the voltage resources used to suppress harmonics; This means that under the high-speed and high-torque conditions where the motor needs to run smoothly the most, its torque pulsation, vibration noise and harmonic losses will inevitably increase, directly sacrificing the running quality and system efficiency; This design concept runs counter to the goal of pursuing continuous high performance under all operating conditions.

[0008] 3. Regarding the control algorithm, it suffers from the dual drawbacks of high real-time computational burden and low engineering adaptability: This scheme requires complex analytical formula calculations dependent on real-time angles within each control cycle, posing a significant challenge to processor performance. Furthermore, this "rigid" algorithm based on fixed mathematical formulas lacks flexibility; when the motor itself or system operating conditions change, it is difficult to conveniently adjust and optimize parameters, resulting in higher difficulty and debugging costs for its engineering application.

[0009] The information disclosed in the background section above is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0010] The purpose of this invention is to provide a fault-tolerant control system and method for a multiphase permanent magnet synchronous motor to solve the problems mentioned in the background art.

[0011] To achieve the above objectives, the present invention provides the following technical solution:

[0012] A fault-tolerant control method for a multiphase permanent magnet synchronous motor, comprising the following steps:

[0013] S1. Real-time acquisition of current signals of each phase in the multiphase permanent magnet synchronous motor, and determination based on the current signals whether at least one phase winding in the multiphase permanent magnet synchronous motor has an open circuit fault, and identification of the faulty phase;

[0014] S2. When the open-circuit fault is detected, the following cooperative operation is performed:

[0015] Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault.

[0016] Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated;

[0017] S3. In the current control loop, a harmonic suppression controller is used to actively suppress the specific harmonic current generated by the asymmetrical operation of the motor due to the open circuit fault, so as to reduce the total current harmonic distortion rate.

[0018] S4. Combining the recalculated current reference command with the output of the harmonic suppression controller, a new set of drive signals is generated through a pulse width modulation algorithm to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby achieving stable and efficient operation of the multiphase permanent magnet synchronous motor under fault conditions.

[0019] Preferably, step S1 specifically includes:

[0020] After acquiring the real-time current of each phase, a current and vector magnitude value characterizing the static imbalance of the three-phase current of the motor, and a maximum phase current fluctuation rate characterizing the instantaneous drastic change of the current are calculated in parallel. Based on preset weights, the current and vector magnitude value and the maximum phase current fluctuation rate are weighted and fused to generate a single fault confidence index.

[0021] Preferably, step S1 further includes:

[0022] A preset fault determination threshold ThF is used. During the current control cycle, the real-time calculated fault confidence index is compared with the fault determination threshold to determine whether an open-circuit fault exists and to identify the specific faulty phase. The fault confidence index is denoted as F, and the specific comparison content is as follows:

[0023] When F is greater than or equal to the fault determination threshold ThF, the value of the "fault confirmation counter" is incremented by 1.

[0024] When F is less than the fault determination threshold ThF, the "fault confirmation counter" is immediately cleared to zero.

[0025] When the fault confirmation flag is generated, the current amplitude and current fluctuation rate of each phase are evaluated in parallel; the phase with a current amplitude less than the preset zero current threshold and the largest current fluctuation rate among all phases is identified as the fault phase, and its phase identifier is output.

[0026] Preferably, step S2 specifically includes:

[0027] After reconstructing the motor mathematical model, the current reference command is recalculated in the following manner:

[0028] First, a set of basic current compensation values ​​are calculated based on the reconstructed mathematical model of the motor to maintain a constant average torque; then, an additional harmonic current component is calculated in parallel to actively suppress torque ripple.

[0029] Preferably, step S2 further includes:

[0030] Finally, the basic current compensation value and the additional harmonic current component are vector-superimposed, and an adjustable torque ripple suppression factor is introduced to dynamically optimize the superposition result, thereby generating a set of final current reference commands that can both guarantee average torque output and actively suppress torque ripple.

[0031] Preferably, step S3 specifically includes:

[0032] The harmonic suppression controller used is an adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller first dynamically selects and sets one or more target resonant frequencies from a preset harmonic frequency library according to the identified fault mode;

[0033] The bandwidth adaptive factor is calculated online based on the real-time operating status of the motor, and the bandwidth adaptive factor is applied to the harmonic suppression controller to adjust the resonant bandwidth of the harmonic suppression controller at each target resonant frequency in real time.

[0034] Preferably, step S3 further includes:

[0035] The difference between the current reference command and the actual current value is obtained to form a current error signal; the current error signal is input to the adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller performs high-gain amplification processing on the current error signal at one or more target resonant frequencies based on the current error signal and the dynamically adjusted resonant bandwidth to generate the harmonic compensation voltage signal.

[0036] Preferably, step S4 specifically includes:

[0037] The recalculated current reference command is vector-superimposed with the output of the harmonic suppression controller to form an initial modulation voltage command;

[0038] Specifically, a bus voltage utilization factor is further calculated, and the dynamic zero-sequence voltage component is calculated based on the bus voltage utilization factor.

[0039] Preferably, step S4 further includes:

[0040] The dynamic zero-sequence voltage component is injected into the initial modulation voltage command to form the final modulation voltage command; and a pulse width modulation algorithm is used to convert the final modulation voltage command into a drive signal for driving the multiphase inverter.

[0041] A fault-tolerant control system for a multiphase permanent magnet synchronous motor includes the following modules:

[0042] The operation status monitoring module is used to acquire the phase current signals of the multiphase permanent magnet synchronous motor in real time, and to determine whether at least one phase winding has an open circuit fault based on the phase current signals, and to identify the faulty phase.

[0043] When the fault-tolerant control reconfiguration module determines that the open-circuit fault exists, it performs the following cooperative operation:

[0044] Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault.

[0045] Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated;

[0046] An asymmetric harmonic suppression module is used in the current control loop to actively suppress specific harmonic currents generated by the asymmetric operation of the motor due to the open circuit fault, thereby reducing the total current harmonic distortion rate.

[0047] The drive signal generation module, in conjunction with the recalculated current reference command and the output of the harmonic suppression controller, generates a new set of drive signals through a pulse width modulation algorithm. These signals are used to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby enabling the multiphase permanent magnet synchronous motor to operate smoothly and efficiently under fault conditions.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] This invention overcomes the passive and idealistic reliability limitations of existing technologies. Specifically, it constructs an active and robust fault diagnosis and confirmation closed loop by parallelly calculating the current and vector magnitude representing static imbalance and the maximum phase current fluctuation rate representing instantaneous changes, and then weighting and fusing them to generate a single fault confidence index F. This F is then compared with the fault confirmation counter and the fault judgment threshold ThF, avoiding reliance on the ideal premise that the fault has been accurately identified. At the same time, the subsequent control reconfiguration does not rely on a single voltage limiting calculation, but is achieved through current command recalculation and closed-loop harmonic suppression, reducing the sensitivity to the accuracy of motor parameters, thereby improving the stability and overall reliability of the system under real operating conditions.

[0050] This invention also solves the problem of the compromise strategy of "downgrading to maintain operation" in the performance after fault tolerance in the prior art; this invention does not simply reduce harmonic suppression resources, but introduces an adjustable torque ripple suppression factor in the current reference command recalculation step, and dynamically optimizes the basic current compensation value used to maintain average torque and the additional harmonic current component used to suppress ripple, thus achieving the dual goal of actively pursuing torque stability while ensuring basic power; furthermore, in the drive signal generation module, by calculating the bus voltage utilization factor and injecting dynamic zero-sequence voltage components, the bus voltage utilization is significantly improved, ensuring a larger voltage margin for harmonic suppression and output torque, and ensuring that the motor can still maintain high-performance operation quality with low torque ripple and low vibration noise under high speed and high torque conditions;

[0051] This invention also employs an asymmetric harmonic suppression module. Its core is not complex analytical calculation, but rather an online adjustment of the final resonant bandwidth of the adaptive bandwidth quasi-proportional resonant controller using a bandwidth broadening factor calculated based on a weighted fusion of the normalized torque command change rate and the normalized harmonic current error. This lightweight adaptive law-based adjustment method significantly reduces the real-time computational burden on the processor. Furthermore, by adjusting the basic resonant bandwidth, maximum additional bandwidth, dynamic weights, and error weights, it endows the algorithm with extremely high flexibility and engineering adaptability, enabling convenient matching with different motor bodies and system operating conditions, significantly reducing the difficulty and debugging cost of engineering applications. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0053] Figure 2 This is a schematic diagram of the overall system framework of the present invention. Detailed Implementation

[0054] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0056] Example 1:

[0057] Please see Figure 1 This invention provides a technical solution: a fault-tolerant control method for a multiphase permanent magnet synchronous motor, the specific steps of which include:

[0058] S1. Real-time acquisition of current signals of each phase in the multiphase permanent magnet synchronous motor, and determination based on the current signals whether at least one phase winding in the multiphase permanent magnet synchronous motor has an open circuit fault, and identification of the faulty phase;

[0059] S2. When the open-circuit fault is detected, the following cooperative operation is performed:

[0060] Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault.

[0061] Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated;

[0062] S3. In the current control loop, a harmonic suppression controller is used to actively suppress the specific harmonic current generated by the asymmetrical operation of the motor due to the open circuit fault, so as to reduce the total current harmonic distortion rate.

[0063] S4. Combining the recalculated current reference command with the output of the harmonic suppression controller, a new set of drive signals is generated through a pulse width modulation algorithm to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby achieving stable and efficient operation of the multiphase permanent magnet synchronous motor under fault conditions.

[0064] Furthermore, existing fault-tolerant control technologies employ isolated processing or pre-defined, insufficiently refined degradation strategies. This technical solution proposes a "globally collaborative, dynamically optimized" fault-tolerant control mechanism, specifically manifested in the following three aspects:

[0065] 1. Differences in control ideologies:

[0066] That is, from "static compensation" to "dynamic model reconstruction", existing technologies usually adopt fixed compensation strategies after detecting a fault; for example, lookup table method or simple current vector rotation, forcibly assign some pre-calculated current values ​​to the remaining phases. Although this method is simple, it does not fundamentally change the controller, causing the controller to still use the full-order mathematical model before the fault to control the already faulty motor, resulting in poor control accuracy and slow dynamic response.

[0067] The S2 in this technical solution adopts a more fundamental "dynamic model reconstruction". The difference from the prior art is that after the faulty phase is identified, a reduced-order mathematical model is dynamically reconstructed based on the system topology structure composed of the remaining normal operating phases to describe the motor operating state after the fault, so as to replace the full-order mathematical model before the fault; and based on the reduced-order mathematical model, the current reference command of the normal operating phase is recalculated.

[0068] 2. Differences in control objectives:

[0069] Unlike existing technologies that limit fault-tolerant control objectives to maintaining average torque output, this invention establishes harmonic suppression under asymmetrical operation as a core control objective that runs parallel to average torque control. Harmonic currents introduced by system asymmetry after a fault are the root cause of torque pulsation, vibration noise, and additional losses.

[0070] To this end, after generating the basic torque current command, the present invention specifically sets up a harmonic suppression feedforward compensation module. First, based on the known fault phase location and the current motor operating state, the amplitude and phase of a specific harmonic component that must exist in the remaining normal phase current are accurately identified or analyzed online. Then, a compensation command with the same amplitude and opposite phase as the harmonic component is generated, and it is vector-superimposed with the original torque current command to form the final composite current reference command.

[0071] By employing this precise hedging method, the present invention achieves source suppression of critical harmonics without affecting the average torque output, thereby elevating the control effect from a simple "torque maintenance" to a new level of "high-quality torque output," significantly improving the operating quality of the motor after a fault.

[0072] 3. Differences in control strategies: From "isolated control" to "cooperative drive":

[0073] In existing technologies, fault-tolerant compensation and conventional current control are often two relatively independent components, lacking deep coupling. The core of this technology lies in "synergy," where the generation of the drive signal simultaneously combines the "current command recalculated to maintain torque" and the "compensation signal generated to suppress harmonics." This means that when the controller issues each PWM pulse, it simultaneously considers two objectives: "how much force to output" and "how to eliminate accompanying vibrations and noise." This is a highly synergistic, multi-objective optimization control strategy that ensures the perfection of the final output.

[0074] Step S1 specifically includes:

[0075] After acquiring the real-time current of each phase, a current and vector magnitude value characterizing the static imbalance of the three-phase current of the motor, and a maximum phase current fluctuation rate characterizing the instantaneous drastic change of the current are calculated in parallel. Based on preset weights, the current and vector magnitude value and the maximum phase current fluctuation rate are weighted and fused to generate a single fault confidence index.

[0076] Step S1 specifically also includes:

[0077] A preset fault determination threshold ThF is used. During the current control cycle, the real-time calculated fault confidence index is compared with the fault determination threshold to determine whether an open-circuit fault exists and to identify the specific faulty phase. The fault confidence index is denoted as F, and the specific comparison content is as follows:

[0078] When F is greater than or equal to the fault determination threshold ThF, the value of the "fault confirmation counter" is incremented by 1.

[0079] When F is less than the fault determination threshold ThF, the "fault confirmation counter" is immediately cleared to zero.

[0080] This embodies the principle of "continuous satisfaction," where any interruption will restart the counting process.

[0081] When the fault confirmation flag is generated, the current amplitude and current fluctuation rate of each phase are evaluated in parallel; the phase with a current amplitude less than the preset zero current threshold and the largest current fluctuation rate among all phases is identified as the fault phase, and its phase identifier is output.

[0082] Furthermore, to achieve step S1 above, the specific calculation process is broken down as follows:

[0083] S1.1. During each control cycle, the instantaneous values ​​of the current in each phase of the multiphase permanent magnet synchronous motor are collected by current sensors, followed by calculation of static unbalance characteristic quantities. Specifically:

[0084] S1.2 By vector summing the real-time current values ​​of all phases in a stationary coordinate system, a composite current vector is obtained; the "current and vector magnitude" is the magnitude of the composite current vector, which is close to zero when the motor is running normally, and deviates significantly from zero after an open circuit fault occurs.

[0085] Divide the calculated "current and vector magnitude" by the rated phase current amplitude of the motor to obtain a dimensionless "normalized current vector magnitude", and map its value range to the [0,1] interval;

[0086] S1.3 Next, the dynamic transient characteristic quantities are calculated, specifically:

[0087] For each phase, the absolute value of the difference between the current value in the current control cycle and the current value in the previous control cycle is calculated to obtain the current change rate of each phase. The "maximum phase current fluctuation rate" is the maximum value among all phase current change rates. This value is controlled by the controller commands during normal operation and remains within a small range. However, a spike far exceeding the normal range will occur at the moment an open circuit fault occurs.

[0088] The calculated "maximum phase current fluctuation rate" is divided by a preset "maximum allowable normal fluctuation rate," which is determined based on the current change rate under the maximum acceleration and deceleration conditions during normal motor operation. After processing, a dimensionless "normalized current fluctuation rate" is obtained, whose value range is also mapped to the [0,1] interval, with any value exceeding 1 treated as 1.

[0089] S1.4 Weighted fusion calculation of composite fault indicators:

[0090] A weighted average method is used to fuse the two normalized feature quantities. This step requires determining two weighting coefficients: a "static unbalanced weight" and a "dynamic transient weight." The sum of these two weighting coefficients is 1, and their specific values ​​can be determined using principal component analysis (PCA) based on offline simulation or experimental data to reflect the relative importance of the two feature quantities in fault diagnosis. For example, to emphasize detection speed, the value of the "dynamic transient weight" can be appropriately increased.

[0091] The final calculation method of the "fault confidence index" can be expressed as follows: multiply the "normalized current vector magnitude" by the "static imbalance weight", and then add the "normalized current volatility" multiplied by the "dynamic transient weight". The sum of the two is the "fault confidence index".

[0092] S1.5 Fault Determination and Location. The calculated "fault confidence index" is compared with the preset "fault determination threshold". If the index exceeds the threshold for several consecutive control cycles, an open-circuit fault is determined to have occurred in the system. Simultaneously, based on the principle that the "maximum phase current fluctuation rate" is the highest and its corresponding phase current amplitude approaches zero, the specific faulty phase is identified and located; in this embodiment, the "fault determination threshold" is initially set to 0.8.

[0093] To quantitatively illustrate the beneficial effects of step S1, the following symbols are defined:

[0094] F is defined as the fault confidence index of the final output; M is defined as the normalized current and vector magnitude.

[0095] R is defined as the normalized maximum phase current ripple rate; wM is defined as the static imbalance weight; wR is defined as the dynamic transient weight.

[0096] The calculation logic of the fault confidence index F is: F equals wM multiplied by M plus wR multiplied by R. The formula is a linear weighted sum. The logical basis is that the occurrence of a fault will leave evidence in two dimensions at the same time: one is the destruction of the steady-state symmetry of the system represented by M, and the other is the instantaneous change of the signal represented by R. Combining the two can form a more reliable criterion.

[0097] Among them, M and R are obtained by calculating and normalizing the phase current data acquired in real time, and the calculation method has been described in detail in the second paragraph; their initial sources are based on physical laws: M comes from Kirchhoff's current law in circuit theory, that is, in a star connection, the vector sum of the ideal three-phase currents is zero; R comes from the concept of derivative in mathematics, which is used to capture the rate of change of the signal.

[0098] wM and wR are preset parameters whose sum is 1. For example, wM = 0.4 and wR = 0.6 can be set. They are obtained by training a large amount of normal and fault condition data using offline optimization algorithms, including particle swarm optimization, with the objective function being to minimize the false alarm rate and the false alarm rate, in order to find the optimal weight combination.

[0099] Furthermore, after normalization, M and R are both dimensionless pure numbers, and wM and wR are also dimensionless weighting coefficients. Therefore, the final F is also a dimensionless exponent, with consistent physical dimensions and logical consistency.

[0100] Since the ranges of M and R are both [0,1], and wM+wR=1, the output range of F is limited to the range of [0,1]. The purpose is to map complex fault states into an intuitive confidence index, which makes it easier to set clear judgment thresholds.

[0101] When F = 0, it is determined that there is no fault.

[0102] When the output F approaches 0, it indicates that the system is operating in a healthy and stable state, and that both M and R are approaching 0. Specifically, M approaching 0 means that the current in each phase is increasingly following a symmetrical distribution, the vector sum is close to zero, and the system has good static balance. R approaching 0 means that the changes in the current in each phase are becoming smoother, without any unexpected drastic fluctuations, and the system has good dynamic stability. At this time, the monitoring module continuously confirms the health status of the system.

[0103] When F=1, it is determined to be a definite fault, and fault-tolerant control is immediately triggered;

[0104] As the output F approaches 1, the trend indicates that the characteristics of an open-circuit fault are becoming more and more obvious; it also indicates that at least one of M and R approaches 1, or both of them approach 1 simultaneously; in a typical open-circuit fault scenario, R will first jump to 1 due to the instantaneous drop in current, and then M will also rise rapidly and stabilize near 1 because the fault phase current is always zero, causing the system to remain asymmetrical; the value of F thus rises rapidly to close to 1, realizing a rapid response and reliable locking of the fault, thereby providing a timely and accurate trigger signal for the subsequent fault-tolerant control reconfiguration module.

[0105] Step S2 specifically includes:

[0106] After reconstructing the motor mathematical model, the current reference command is recalculated in the following manner:

[0107] First, a set of basic current compensation values ​​are calculated based on the reconstructed mathematical model of the motor to maintain a constant average torque; then, an additional harmonic current component is calculated in parallel to actively suppress torque ripple.

[0108] Step S2 also includes:

[0109] Finally, the basic current compensation value and the additional harmonic current component are vector-superimposed, and an adjustable torque ripple suppression factor is introduced to dynamically optimize the superposition result, thereby generating a set of final current reference commands that can both guarantee average torque output and actively suppress torque ripple.

[0110] Furthermore, to achieve step S2 above, the specific calculation process is broken down as follows:

[0111] S2.1 Upon receiving a fault phase signal, immediately update the Clarke transformation matrix in the fault-tolerant control system; directly reduce the original N×N dimensional transformation matrix to a new (Nk)×(Nk) dimensional matrix to establish a mathematical model reflecting the current topology of the remaining normally operating phases; where N is the total number of phases and k is the number of fault phases;

[0112] Among them, the Clark transformation matrix aims to reduce the dimensionality and decouple the control problem of a multiphase AC system by simplifying complex time-varying phasors through mathematical mapping, thereby facilitating the realization of high-performance vector control;

[0113] S2.2 Using the new (Nk)×(Nk) dimensional matrix in S2.1, the DC current command of the dq axis output by the torque loop, i.e. the desired torque and flux linkage components, is transformed into a set of AC current commands that vary with time for each of the remaining normal operating phases. This set of AC current commands consisting of all the remaining normal phases is defined as the "base current compensation value", the goal of which is to ensure that the average electromagnetic torque of the motor after the fault remains consistent with that before the fault.

[0114] S2.3. Based on the number of motor phases and fault mode, determine the harmonic order that causes the main torque pulsation; derive the reconstructed torque equation based on the reconstructed motor mathematical model to describe the relationship between the current and electromagnetic torque of the remaining normally operating phases.

[0115] Specifically, the reconstructed torque equation is obtained by applying the general electromagnetic torque principle to a new electrical model of a motor that has been reduced in dimension and asymmetric due to faults, thus obtaining an analytical expression.

[0116] During the process, the "basic current compensation value" is used as input and substituted into the reconstructed torque equation for forward analysis to quantitatively identify the specific harmonic torque pulsation components that are inevitably generated due to the asymmetry of the motor structure, and to determine their harmonic order, amplitude and phase; the specific harmonics, taking the single-phase open circuit fault of a five-phase motor as an example, are mainly the 2nd and 4th harmonics.

[0117] Finally, by solving the reconstructed torque equation in reverse, it calculates the specific subharmonic current required to generate a compensating torque that is equal in magnitude and opposite in phase to the harmonic torque pulsation component, which is defined as the additional harmonic current component.

[0118] S2.4 The torque ripple suppression factor is a dimensionless adjustable parameter with a value range between [0,1], used to balance the two objectives of maintaining average torque and suppressing torque ripple. The value of the torque ripple suppression factor can be preset by the user according to the application requirements, that is, to pursue extreme stability or to take efficiency into account, or to be dynamically adjusted by an adaptive algorithm according to real-time speed, load and other operating conditions.

[0119] S2.5 The calculation process for the final current reference command is as follows:

[0120] The additional harmonic current component is multiplied by the torque ripple suppression factor, and the result is vector-sumped with the base current compensation value; this final vector sum is the "final current reference command" sent to each normal operating phase of the subsequent current control loop.

[0121] Furthermore, to quantitatively illustrate the beneficial effects of this scheme, the following symbols are defined:

[0122] Iref is defined as the final current reference instruction generated at the end;

[0123] Ibase is defined as the base current compensation value;

[0124] Iharm is defined as the additional harmonic current component;

[0125] Ktps is defined as the Torque-Pulsation-Suppression Factor;

[0126] The calculation formula for Iref is a dynamic weighted superposition; its logical basis is that the ideal current after a fault consists of two parts: one part Ibase is used to generate a constant average torque, and the other part Iharm is used to offset the pulsating torque caused by the asymmetric structure, while Ktps is used to control the degree of pulsation suppression function.

[0127] Furthermore, Ibase and Iharm are obtained through mathematical operations on the motor model, with their initial sources being the electromagnetic torque equation in motor mechanics and harmonic analysis in multiphase motor theory; the calculation of Ibase depends on the reconstructed coordinate transformation; the calculation of Iharm depends on the Fourier decomposition of the torque equation after the fault; the lower-level parameters in step S2 are all based on known motor parameters and control commands.

[0128] Ktps is a configurable control parameter with a value range of [0,1]. In this embodiment, Ktps = 1 is set for precision instrument drives where stability is extremely important. In applications where some pulsation is permissible but efficiency is of greater concern, Ktps = 0.5.

[0129] The physical dimensions of Iref, Ibase, and Iharm are all electric current; Ktps is a dimensionless coefficient; therefore, the physical dimensions of both sides of the formula are consistent, namely electric current, which conforms to the laws of physics.

[0130] When Ktps is 0, it means that torque ripple suppression is not performed at all. In this case, Iref equals Ibase, and the control strategy degenerates into the traditional fault-tolerant method of "maintaining only the average torque".

[0131] When Ktps is set to 1, it indicates that maximum torque ripple suppression is performed. In this case, the controller will inject all calculated harmonic components to achieve the smoothest torque output.

[0132] Step S3 specifically includes:

[0133] The harmonic suppression controller used is an adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller first dynamically selects and sets one or more target resonant frequencies from a preset harmonic frequency library according to the identified fault mode;

[0134] The bandwidth adaptive factor is calculated online based on the real-time operating status of the motor, and the bandwidth adaptive factor is applied to the harmonic suppression controller to adjust the resonant bandwidth of the harmonic suppression controller at each target resonant frequency in real time.

[0135] Step S3 specifically also includes:

[0136] The difference between the current reference command and the actual current value is obtained to form a current error signal; the current error signal is input to the adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller performs high-gain amplification processing on the current error signal at one or more target resonant frequencies based on the current error signal and the dynamically adjusted resonant bandwidth to generate the harmonic compensation voltage signal.

[0137] Furthermore, the specific implementation process for step S3 above is broken down as follows:

[0138] S3.1 The method for setting the target resonant frequency and obtaining the current error is as follows: Based on the identified fault mode, one or more "target resonant frequencies" are dynamically selected and set from the preset "fault mode-harmonic frequency" mapping table. At the same time, the "final current reference command" output by the fault-tolerant control reconfiguration module and the actual current feedback value are obtained, and the difference between the two is calculated to form the "current error signal".

[0139] S3.2 The process of calculating the characteristic quantities required for the bandwidth broadening factor is as follows:

[0140] Calculate the "torque command change rate": Obtain the torque command value of the current control cycle and the torque command value of the previous control cycle, calculate the absolute value of the difference between the two, and obtain the "torque command change rate". This value proactively reflects the dynamic trend of the motor.

[0141] Calculate the "harmonic current error": Perform a fast Fourier transform on the "current error signal" to extract the error component amplitude at each "target resonant frequency", and take the maximum value as the "harmonic current error". This value reflects the actual effect of harmonic suppression.

[0142] Normalization: Divide the "torque command change rate" by a preset "maximum permissible torque change rate" to obtain the "normalized torque command change rate"; divide the "harmonic current error" by a preset "maximum permissible harmonic error" to obtain the "normalized harmonic current error". Both normalized values ​​are mapped to the [0,1] interval.

[0143] S3.3 The weighted fusion calculation process for the bandwidth broadening factor is as follows:

[0144] Determine the weighting coefficients: The weighted average method is used to fuse the two normalized characteristic quantities mentioned above; it is necessary to determine the "dynamic weight" and the "error weight", the sum of which is 1; their values ​​can be determined based on a large amount of operating data using principal component analysis to reflect the relative importance of "dynamic feedforward" and "error feedback" in determining the bandwidth adjustment strategy.

[0145] The final calculation method for the "bandwidth expansion factor" is as follows: multiply the "normalized torque command change rate" by the "dynamic weight", and then add the "normalized harmonic current error" multiplied by the "error weight". The sum of the two is the "bandwidth expansion factor".

[0146] S3.4. The "bandwidth widening factor" is used to adjust the resonant bandwidth in real time. The specific calculation process is as follows: multiply a preset "maximum additional bandwidth" by the "bandwidth widening factor", and add the result to a preset "basic resonant bandwidth" to obtain the "final resonant bandwidth". Finally, the "current error signal" is input to the adaptive bandwidth quasi-proportional resonant controller with this "final resonant bandwidth" for high gain amplification to generate the "harmonic compensation voltage signal".

[0147] Furthermore, to quantitatively illustrate the beneficial effects of this scheme, the following symbols are defined:

[0148] Kbw represents the final calculated bandwidth-widening factor;

[0149] ΔT represents the normalized rate of change of torque command;

[0150] Eh represents the normalized harmonic current error;

[0151] wT represents the dynamic weight; wE represents the error weight;

[0152] BW final This is represented as the final resonant bandwidth of the controller;

[0153] BW base This is represented as the preset fundamental resonant bandwidth;

[0154] BWadd max This represents the preset maximum additional bandwidth;

[0155] In step S3, there are two key calculation formulas, as follows:

[0156] Formula 1: Kbw equals wT multiplied by ΔT plus wE multiplied by Eh;

[0157] Formula 2: BW final Equal to BW base Add BWadd max Multiply by Kbw;

[0158] Formula 1 is a linear weighted sum, its initial source being the weighted scoring model in data fusion and decision theory. It merges two physically distinct indicators, dynamic look-ahead and steady-state feedback, into a single Kbw that characterizes the system's "instability."

[0159] Formula 2 is a linear mapping function whose initial source is the parameter adaptive law design in control theory; it utilizes Kbw to map the key controller parameter BW. final Make smooth, linear adjustments within a preset, reasonable range.

[0160] The entire process is logically clear: perceive the state > comprehensive evaluation > adjust strategy.

[0161] Parameter acquisition and source:

[0162] ΔT and Eh are obtained by calculating the torque command inside the controller and the current error acquired externally, respectively. Their sources are control system theory and signal processing theory.

[0163] wT, wE, BW base and BWadd max All parameters are preset based on motor characteristics and control requirements; for a 1kW five-phase motor, this embodiment will use BW base Set to 5 rad / s, BWadd max Set to 20 rad / s, and set wT = 0.4 and wE = 0.6.

[0164] Dimensionality Consistency: In Formula 1, all inputs are dimensionless numbers, therefore Kbw is also a dimensionless number; in Formula 2, BW base and BWadd max The dimension of is angular frequency (rad / s), while Kbw is dimensionless, therefore BW final The dimension of is also angular frequency (rad / s), which is consistent with the dimensions and conforms to the laws of physics;

[0165] Since the ranges of ΔT and Eh are both [0,1], and wT+wE=1, the output range of Kbw is strictly limited to the range [0,1]. This factor, as the core driving force for bandwidth adjustment, directly reflects the system's demand for wide bandwidth.

[0166] Step S4 specifically includes:

[0167] The recalculated current reference command is vector-superimposed with the output of the harmonic suppression controller to form an initial modulation voltage command;

[0168] Specifically, a bus voltage utilization factor is further calculated, and the dynamic zero-sequence voltage component is calculated based on the bus voltage utilization factor.

[0169] Step S4 specifically also includes:

[0170] The dynamic zero-sequence voltage component is injected into the initial modulation voltage command to form the final modulation voltage command; and a pulse width modulation algorithm is used to convert the final modulation voltage command into a drive signal for driving the multiphase inverter.

[0171] Furthermore, to achieve the generation of the drive signal in step S4 above, the specific implementation process is broken down as follows:

[0172] S4.1: Synthesis of initial modulation voltage command; Obtain the "final current reference command" output in step S2 and send it to the main current regulator to obtain the main regulation voltage; At the same time, obtain the "harmonic compensation voltage signal" output in step S3; Add the main regulation voltage and the "harmonic compensation voltage signal" vectorively to obtain the "initial modulation voltage command" for each normal operating phase.

[0173] S4.2 Calculate the "maximum phase voltage demand": In the current control cycle, traverse the "initial modulation voltage command" of all normally operating phases and find the maximum value of its amplitude. This value is defined as the "maximum phase voltage demand".

[0174] The "bus voltage utilization factor" is calculated by dividing the "maximum phase voltage demand" by a preset "maximum linear modulation voltage amplitude". The result is the "bus voltage utilization factor", which reflects the degree to which the current voltage command approximates the bus voltage.

[0175] S4.3 Obtain the voltage command for each phase: Obtain the instantaneous value of the "initial modulation voltage command" for each normally operating phase;

[0176] At the current sampling moment, iterate through the instantaneous voltage command values ​​of all normally operating phases, find the maximum and minimum values, and define them as the "maximum phase voltage command" and the "minimum phase voltage command" respectively.

[0177] The "dynamic zero-sequence voltage component" is calculated as follows: add the "maximum phase voltage command" and the "minimum phase voltage command", and then multiply the sum by "-0.5". The result is the "dynamic zero-sequence voltage component". This calculation method corresponds to the minimum-maximum injection method, which aims to reduce the overall amplitude of the modulated wave.

[0178] S4.4, Final modulation voltage command generation and PWM drive signal output: The "dynamic zero-sequence voltage component" calculated in step three is added to the "initial modulation voltage command" of each phase to obtain the "final modulation voltage command" of each phase; then, this set of "final modulation voltage commands" is sent to the pulse width modulator and compared with the carrier signal to generate a set of PWM drive signals to drive the power switching transistors of each phase inverter.

[0179] Furthermore, to quantitatively illustrate the beneficial effects of this scheme, the following symbols are defined:

[0180] Vmo dfinal Defined as the final modulation voltage command;

[0181] Vmo dinit Defined as the initial modulation voltage command;

[0182] Vze is defined as the dynamic zero-sequence voltage component;

[0183] Kdcu is defined as the DC-bus-Utilization-Factor;

[0184] Vreq max Defined as the maximum phase voltage requirement;

[0185] Vlin max Defined as the maximum linear modulation voltage amplitude;

[0186] Vcmd max Defined as the maximum phase voltage command;

[0187] Vcmd min Defined as minimum phase voltage command;

[0188] Furthermore, S4 contains the following three formulas, as detailed below:

[0189] Formula 1:

[0190] Formula 2: Vze=-0.5×(Vcmd max +Vcmd min );

[0191] Formula 3: Vmo dfinal =Vmo dinit +Vze;

[0192] Logical Relationships and Derivation: Formula 1 is the state assessment, used to quantify the margin of the current voltage command; Formula 2 is the core optimization algorithm, initially derived from advanced PWM modulation strategies in power electronics, particularly the zero-sequence component injection concept in Space Vector Pulse Width Modulation (SVPWM), which calculates a voltage component that shifts the entire modulated wave towards zero; Formula 3 is the execution injection, superimposing the optimized amount onto the original command. The entire process achieves a complete closed loop from "assessment" to "optimization" to "execution," demonstrating rigorous logic.

[0193] Among them, Vreq max Vcmd max and Vcmd min All are obtained through real-time comparison and calculation of the "initial modulation voltage command" calculated internally by the controller; Vlin max It is a system constant determined by the DC bus voltage of the inverter, and its origin is the definition of the linear operating region of the inverter in power electronics.

[0194] In Formula 1, the numerator and denominator are both voltage V, so Kdcu is a dimensionless number; in Formulas 2 and 3, the dimension of all terms is voltage V, and the dimensions of the left and right sides of the formulas are consistent.

[0195] Theoretically, the output range of Kdcu is [0, +∞), but in the actual linear modulation region, its effective range is [0, 1]. As an indicator of system voltage margin, it directly reflects whether and to what extent modulation optimization is needed.

[0196] When Kdcu is 0, it indicates that the motor is stationary and there is no voltage command.

[0197] In this invention, Kdcu is mainly used for state monitoring, while zero-sequence injection (Formula 2) is always executed. Its injection amount is determined by the voltage command itself, rather than being directly controlled by Kdcu.

[0198] Example 2

[0199] Please see Figure 2 A fault-tolerant control system for a multiphase permanent magnet synchronous motor includes the following modules:

[0200] The operation status monitoring module is used to acquire the phase current signals of the multiphase permanent magnet synchronous motor in real time, and to determine whether at least one phase winding has an open circuit fault based on the phase current signals, and to identify the faulty phase.

[0201] When the fault-tolerant control reconfiguration module determines that the open-circuit fault exists, it performs the following cooperative operation:

[0202] Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault.

[0203] Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated;

[0204] An asymmetric harmonic suppression module is used in the current control loop to actively suppress specific harmonic currents generated by the asymmetric operation of the motor due to the open circuit fault, thereby reducing the total current harmonic distortion rate.

[0205] The drive signal generation module, in conjunction with the recalculated current reference command and the output of the harmonic suppression controller, generates a new set of drive signals through a pulse width modulation algorithm. These signals are used to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby enabling the multiphase permanent magnet synchronous motor to operate smoothly and efficiently under fault conditions.

[0206] In this embodiment, to verify the effectiveness of the adaptive bandwidth harmonic suppression and dynamic zero-sequence voltage injection coordinated control in the fault-tolerant control method for multiphase permanent magnet synchronous motors described in this invention, the following example is designed and implemented:

[0207] The experimental platform was set up as follows:

[0208] A five-phase permanent magnet synchronous motor with a rated power of 1.5kW, a rated voltage of 311V, and a rated speed of 1500rpm is used as the controlled object.

[0209] The drive system is a five-phase voltage source inverter with integrated IGBT power modules. The DC bus voltage Vdc is stabilized at 320V. When the inverter operates in the linear modulation region and uses the SVPWM modulation algorithm, its maximum output linear modulation voltage amplitude Vlin is [value missing]. max Vdc / √3≈184.7V;

[0210] The control core adopts a fast prototype controller based on the TITMS320F28377D digital signal processor, which has sufficient computing power to execute the algorithm described in this invention;

[0211] The measuring equipment included a Tektronix high-precision current probe, a Yokogawa power analyzer, and a high-bandwidth oscilloscope. In the experiment, the open-circuit fault of phase A winding was first simulated by software. After the operating status monitoring module accurately identified and isolated the faulty phase, the system entered the fault-tolerant operation mode.

[0212] To compare the effects, two control strategies were set up, including a comparative scheme and an inventive scheme:

[0213] Comparison scheme: The traditional fixed bandwidth quasi-proportional resonant (QPR) controller is used, with its resonant bandwidth fixed at 10 rad / s based on empirical values, and the PWM modulation adopts standard space vector pulse width modulation (SVPWM).

[0214] Invention solution: Complete implementation of the asymmetric harmonic suppression module and drive signal generation module described in this embodiment;

[0215] The experimental procedure is as follows:

[0216] First, with an open-circuit fault in phase A, the motor was stably operated at 1000 rpm and a load torque of 5 N·m using a comparative scheme. At t = 1.0 s, a step torque command was applied to the system, requiring the output torque to jump from 5 N·m to 9 N·m in a very short time; during this dynamic process, key performance indicators were recorded in detail.

[0217] The system was then reset and switched to the invention solution;

[0218] Under identical initial operating conditions (phase A open circuit, 1000 rpm, 5 N·m), the same 9 N·m step torque command is applied at t = 1.0 s; during the dynamic response of the invention, the asymmetric harmonic suppression module inside the controller performs the following operations in real time:

[0219] At the instant the step occurs, a sharp increase in the "torque command change rate" is detected, and its normalized value ΔT instantaneously approaches 1.0; at the same time, due to the dynamic changes of the system, the current tracking error increases, and the extracted "normalized harmonic current error" Eh also increases significantly.

[0220] According to the formula Kbw=wT×ΔT+wE×Eh;

[0221] The preset weights are wT = 0.4 and wE = 0.6.

[0222] The calculated bandwidth expansion factor Kbw reaches a peak.

[0223] According to the formula BW final =BW base +BWadd max ×Kbw;

[0224] Among them, the preset BW base =5 rad / s, BWadd max =20 rad / s;

[0225] The resonant bandwidth was dynamically increased from 5 rad / s in steady state to 5 + 20 * 0.82 = 21.4 rad / s at the peak value.

[0226] This extremely wide bandwidth enables the controller to powerfully and rapidly suppress harmonic current errors generated during dynamic processes; at the same time, as the output torque increases, the voltage command of the drive signal generation module also increases accordingly.

[0227] Under a high load of 9 N·m, the "maximum phase voltage demand" Vreq is calculated. max Reaching 178V makes the calculated value of the "bus voltage utilization factor" Kdcu approximately 178 / 184.7 ≈ 0.964;

[0228] Based on the instantaneous value of the "initial modulation voltage command" for each phase, a significant "dynamic zero-sequence voltage component" Vze is calculated and injected, which effectively expands the linear modulation region and avoids the deterioration of control performance caused by voltage saturation. The data of the entire process is recorded by a power analyzer and an oscilloscope for subsequent comparative analysis.

[0229]

[0230]

[0231] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, min-max-normalization and Z-score standardization.

[0232] The technical solution of this invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random-access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of this invention.

[0233] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0234] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fault-tolerant control method for a multiphase permanent magnet synchronous motor, characterized in that, The specific steps include: S1. Real-time acquisition of current signals of each phase in the multiphase permanent magnet synchronous motor, and determination based on the current signals whether at least one phase winding in the multiphase permanent magnet synchronous motor has an open circuit fault, and identification of the faulty phase; S2. When the open-circuit fault is detected, the following cooperative operation is performed: Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault. Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated; S3. In the current control loop, a harmonic suppression controller is used to actively suppress the specific harmonic current generated by the asymmetrical operation of the motor due to the open circuit fault, so as to reduce the total current harmonic distortion rate. S4. Combining the recalculated current reference command with the output of the harmonic suppression controller, a new set of drive signals is generated through a pulse width modulation algorithm to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby achieving stable and efficient operation of the multiphase permanent magnet synchronous motor under fault conditions.

2. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 1, characterized in that: Step S1 specifically includes: After acquiring the real-time current of each phase, a current and vector magnitude value characterizing the static imbalance of the three-phase current of the motor, and a maximum phase current fluctuation rate characterizing the instantaneous drastic change of the current are calculated in parallel. Based on preset weights, the current and vector magnitude value and the maximum phase current fluctuation rate are weighted and fused to generate a fault confidence index.

3. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 2, characterized in that: Step S1 specifically also includes: A preset fault determination threshold ThF is used. During the current control cycle, the real-time calculated fault confidence index is compared with the fault determination threshold to determine whether an open-circuit fault exists and to identify the specific faulty phase. The fault confidence index is denoted as F, and the specific comparison content is as follows: When F is greater than or equal to the fault determination threshold ThF, the value of the "fault confirmation counter" is incremented by 1. When F is less than the fault determination threshold ThF, the "fault confirmation counter" is immediately cleared to zero. When the fault confirmation flag is generated, the current amplitude and current fluctuation rate of each phase are evaluated in parallel; the phase with a current amplitude less than the preset zero current threshold and the largest current fluctuation rate among all phases is identified as the fault phase, and its phase identifier is output.

4. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 3, characterized in that: Step S2 specifically includes: After reconstructing the motor mathematical model, the current reference command is recalculated in the following manner: Based on the reconstructed mathematical model of the motor, a set of basic current compensation values ​​are calculated to maintain a constant average torque; then, an additional harmonic current component is calculated in parallel to actively suppress torque ripple.

5. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 4, characterized in that: Step S2 also includes: The basic current compensation value and the additional harmonic current component are vector-superimposed, and an adjustable torque ripple suppression factor is introduced to dynamically optimize the superposition result, thereby generating a set of final current reference commands that can both guarantee average torque output and actively suppress torque ripple.

6. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 5, characterized in that: Step S3 specifically includes: The harmonic suppression controller used is an adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller first dynamically selects and sets one or more target resonant frequencies from a preset harmonic frequency library according to the identified fault mode; The bandwidth adaptive factor is calculated online based on the real-time operating status of the motor, and the bandwidth adaptive factor is applied to the harmonic suppression controller to adjust the resonant bandwidth of the harmonic suppression controller at each target resonant frequency in real time.

7. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 6, characterized in that: Step S3 specifically also includes: The difference between the current reference command and the actual current value is obtained to form a current error signal; the current error signal is input to the adaptive bandwidth quasi-proportional resonant controller; the harmonic suppression controller performs high-gain amplification processing on the current error signal at one or more target resonant frequencies based on the current error signal and the dynamically adjusted resonant bandwidth to generate the harmonic compensation voltage signal.

8. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 7, characterized in that: Step S4 specifically includes: The recalculated current reference command is vector-superimposed with the output of the harmonic suppression controller to form an initial modulation voltage command; Specifically, a bus voltage utilization factor is further calculated, and the dynamic zero-sequence voltage component is calculated based on the bus voltage utilization factor.

9. The fault-tolerant control method for a multiphase permanent magnet synchronous motor according to claim 8, characterized in that: Step S4 specifically also includes: The dynamic zero-sequence voltage component is injected into the initial modulation voltage command to form the final modulation voltage command; and a pulse width modulation algorithm is used to convert the final modulation voltage command into a drive signal for driving the multiphase inverter.

10. A fault-tolerant control system for a multiphase permanent magnet synchronous motor, characterized in that: The fault-tolerant control system for a multiphase permanent magnet synchronous motor is used to execute the fault-tolerant control method for a multiphase permanent magnet synchronous motor according to any one of claims 1-9, and includes the following modules: The operation status monitoring module is used to acquire the phase current signals of the multiphase permanent magnet synchronous motor in real time, and to determine whether at least one phase winding has an open circuit fault based on the phase current signals, and to identify the faulty phase. When the fault-tolerant control reconfiguration module determines that the open-circuit fault exists, it performs the following cooperative operation: Dynamic reconfiguration of mathematical model: Based on the remaining normal operating phases, the mathematical model of the motor is dynamically reconfigured to reflect the system topology after the fault. Current reference command recalculation: Based on the reconstructed mathematical model and the preset torque command, the current reference command for the normal operating phase is recalculated and generated. An asymmetric harmonic suppression module is used in the current control loop to actively suppress specific harmonic currents generated by the asymmetric operation of the motor due to the open circuit fault, thereby reducing the total current harmonic distortion rate. The drive signal generation module, in conjunction with the recalculated current reference command and the output of the harmonic suppression controller, generates a new set of drive signals through a pulse width modulation algorithm. These signals are used to drive the multiphase inverter connected to the multiphase permanent magnet synchronous motor, thereby enabling the multiphase permanent magnet synchronous motor to operate smoothly and efficiently under fault conditions.

Images