Motor driving system open-circuit fault diagnosis method

Abstract

The invention discloses an open-circuit fault diagnosis method for a motor driving system, and the method comprises the steps: constructing a fault index of each phase through mathematical transformation, filtering and sliding average processing according to a zero-state conservation equation of a multi-phase motor; according to the characteristic value expression of the fault index under the fault condition, a threshold value judgment method is designed; meanwhile, performing absolute value, average value and filtering processing on the fault index of each phase to construct a fault diagnosis index of an upper switch and a lower switch of each phase; and finally, integrating the upper and lower switch fault diagnosis indexes to obtain each phase fault flag variable, and combining the fault diagnosis table to obtain a fault diagnosis result of the multi-phase motor driver. In a fundamental wave period, the open-circuit fault position of the motor winding and the power device can be rapidly and accurately detected, the universality is high, the diagnosis precision is high, the algorithm is simple, the real-time performance is high, particularly, the robustness to non-sinusoidal current is high, and an effective basis is provided for further fault-tolerant control and system fault processing.

Description

Technical Field

[0001] The present invention belongs to the technical field of motor fault diagnosis, and in particular relates to a method for diagnosing open-circuit faults in a motor drive system. Background Art

[0002] In many applications requiring improved overall system reliability and reduced per-phase power, multiphase motors are recognized as an attractive alternative to traditional three-phase motors. Compared to three-phase motors, multiphase motors offer advantages such as higher reliability, higher power ratings, lower power device rating requirements, and additional degrees of freedom. They can also operate fault-tolerantly even in a phase-loss state in the drive system, leading to their widespread application in areas such as ship propulsion, electric vehicles, and aerospace, which require high fault tolerance and reliability. For this reason, multiphase motor systems require the ability to online identify and locate fault types to ensure reliable and stable operation of the power system. Due to the influence of the operating environment, multiphase motor drive system faults are often concentrated in open windings and open power devices. Therefore, online diagnosis of multiphase motor drive system faults primarily focuses on open windings and open power devices.

[0003] Currently, most of the motor drive open circuit fault diagnosis methods are based on model, signal and data-driven methods. Model-based methods (such as [N.Diao, Y.Zhang, X.Sun, C.Song, W.Wang and H.Zhang,"A Real-Time Open-Circuit Fault Diagnosis Method Based on HybridModel Flux Observer for Voltage Source Inverter Fed Sensorless VectorControlled Drives," in IEEE Transactions on Power Electronics, vol.38, no.2, pp.2539-2551, Feb.2023]) monitor the consistency between the measured output and its observed output, which will make it highly dependent on the motor parameters; data-driven methods (such as [X.Jiang, J.Zhou, K.Wang, S.Yang, Z.Weiand J.Liu,"Open-Circuit Fault Diagnosis Strategy for Novel DFPMM Drive Basedon Fuzzy Logic," in IEEE Transactions on Industrial Electronics, vol.71, no.7, pp.6753-6763, July 2024]) is to extract potential fault features contained in a large amount of historical data through artificial intelligence algorithms. Due to its high dependence on massive data and complex algorithms, the computational cost is expensive.

[0004] For multiphase motor drives, signal-based methods (such as [H. Mesai-Ahmed, I. Jlassi, AJM Cardoso and A. Bentaallah, "Multiple Open-Circuit Faults Diagnosis in Six-Phase Induction Motor Drives Using Stator Current Analysis," in IEEE Transactions on Power Electronics, vol. 37, no. 6, pp. 7275-7285, June 2022]) are relatively simple. They directly utilize the primary behavior of the measured signal for fault diagnosis. However, the multiphase motor open-circuit fault diagnosis algorithm is an extension of the three-phase motor drive fault diagnosis algorithm. As the number of variables processed increases, the calculation becomes more complex and the effect is not very obvious. In addition, as the number of phases in multiphase motor drives increases, the number of objects experiencing open-circuit faults also increases. Therefore, the problem of online diagnosis of multiphase motors needs to be solved and optimized. Summary of the Invention

[0005] In view of the above, the present invention provides a method for diagnosing open-circuit faults in a motor drive system, which can solve the technical problems of complex detection circuits, high detection costs and unclear positioning in the prior art for open-circuit fault diagnosis of multi-phase motors, and can reduce the probability of misdiagnosis of open-circuit faults during no-load and loaded operation of multi-phase motors. It has the function of accurately judging open-circuit faults and clarifying the open-circuit fault points without performing a large number of complex operations.

[0006] A method for diagnosing an open circuit fault in a motor drive system comprises the following steps:

[0007] (1) Acquire the stator current signals of each phase of the multi-phase motor system and construct the initial fault index of each phase based on the zero-state conservation equation of the motor in normal and fault states;

[0008] (2) continuously collecting and counting the initial fault index of each phase;

[0009] (3) filtering and performing sliding average processing on the initial fault index to obtain a fault phase detection index for each phase;

[0010] (4) Calculate and determine the fault phase detection flag of each phase based on the fault phase detection index and the count value and detect the fault phase accordingly;

[0011] (5) For the fault phase, the initial fault index of the phase is processed by absolute value to obtain the fault switch detection index;

[0012] (6) the fault switch detection index is respectively passed through two hysteresis filters to obtain an upper switch tube fault diagnosis index and a lower switch tube fault diagnosis index;

[0013] (7) the upper switch tube fault diagnosis index and the lower switch tube fault diagnosis index are subjected to average value operation to obtain an upper switch tube state monitoring index and a lower switch tube state monitoring index;

[0014] (8) according to the upper switch tube state monitoring index and the lower switch tube state monitoring index, a positioning detection flag of a fault phase is calculated and determined, and the fault position is detected.

[0015] Further, the expression of the initial fault index in the step (1) is as follows:

[0016]

[0017] wherein: for any phase, f is the phase of the phase, P f is the initial fault index of the phase, i k is the instantaneous value of the k-phase stator current signal (unit: ampere), Q f(n) represents 2g adjacent phase types in the counterclockwise direction with the f phase as the symmetry axis, H f(n) represents other phase types except the f phase and Q f(n) g is the maximum integer not greater than n / 4, and n is the number of motor phases.

[0018] Further, the specific implementation mode of the step (2) is that for any phase, the initial fault index of the phase is continuously collected at a fixed sampling frequency, and a counter is used to record the number of continuous samplings of the initial fault index being -1, that is, when the index value is -1, the counter is added by 1; when the index value is not equal to -1, the counter is reset to 0.

[0019] Further, the specific implementation mode of the step (4) is that for any phase, if and J>C th , then SI f =1, indicating that the phase is faulty; if and J≤C th , then SI f =0, indicating that the phase is healthy; wherein is the fault phase detection index of the phase, J is the counting value of the initial fault index of the phase, SI f is the fault phase detection flag of the phase, C th is the set counter threshold, T th is the set fault phase detection threshold.

[0020] Further, the transfer function expressions of the two hysteresis filters in the step (6) are as follows:

[0021]

[0022] wherein: is the fault switch detection index of the fault phase, r fu and r fl are the upper switch tube fault diagnosis index and the lower switch tube fault diagnosis index of the fault phase respectively, and ε is the hysteresis bandwidth of the hysteresis filter.

[0023] Further, the step (8) is specifically implemented as follows: first, the positioning detection index DI of the fault phase is calculated according to the following formula f :

[0024] DI f = F fu + 2*F fl

[0025] wherein: F fu and F fl are the upper switch tube state monitoring index and the lower switch tube state monitoring index of the fault phase respectively.

[0026] When DI f = 1, it indicates that the upper switch tube of the fault phase has an open circuit fault;

[0027] When DI f = 2, it indicates that the lower switch tube of the fault phase has an open circuit fault;

[0028] When DI f = 3, it indicates that the upper switch tube and the lower switch tube of the fault phase both have an open circuit fault, or the winding has a phase breakage fault.

[0029] Further, the stator winding of the multi-phase motor is a single neutral point star connection symmetric structure, that is, the tail ends (or the head ends) of the multiple windings of the motor are connected together to form a common point in the star connection.

[0030] A computer device, comprising a memory and a processor, the memory has a computer program stored therein, and the processor is used to execute the computer program to realize the motor drive system open circuit fault diagnosis method.

[0031] A computer readable storage medium, storing a computer program, the computer program is executed by a processor to realize the motor drive system open circuit fault diagnosis method.

[0032] Based on the above technical solutions, the present application has the following beneficial technical effects:

[0033] 1. The method of the present application does not need to add additional sensors, adopts a non-invasive diagnostic method, and can complete the diagnosis of single or multiple open circuit faults of the inverter switch tube in the multi-phase motor drive system only by collecting the multi-phase stator fault current of the motor output, has low cost, simple calculation, and is easy to realize in real time, and can be widely applied.

[0034] 2. The fault index involved in the present application does not affect each other between phases, so the multiple open circuit fault diagnosis result is obtained separately for each phase fault index, which is convenient for expansion to more phase motor drive systems.

[0035] 3. The zero-state conservation equation involved in the present application is not affected by the operating point, transient process and current distortion, and has natural robustness.

[0036] 4. The present application designs a cascaded diagnosis scheme, in order to further reduce the calculation burden, after obtaining the fault phase detection information, the switch fault diagnosis program is activated, and the robustness is further ensured.

[0037] 5. The diagnosis time of the present application is within one basic current cycle, and the fault switch tube can be accurately and quickly located. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 It is an equivalent circuit structure schematic diagram of the 9-phase motor drive system in the embodiment of the present application.

[0039] Figure 2 It is a schematic diagram of the winding phase distribution of the symmetrical 9-phase brushless DC motor in the embodiment of the present application.

[0040] Figure 3 It is a flowchart of the open circuit fault diagnosis method of the motor drive system of the present application. DETAILED DESCRIPTION

[0041] In order to more specifically describe the present application, the technical solutions of the present application will be described in detail below in combination with the drawings and specific embodiments.

[0042] The present application is suitable for open circuit fault diagnosis of a motor driver, takes the upper and lower switch tubes on the same phase of the driver as a unit base, comprehensively obtains the online real-time diagnosis result of multiple open circuit faults of the driver from the diagnosis results of each phase, collects the phase current of the motor driver through a current sensor, then inputs it into a fault diagnosis algorithm module, finally obtains a fault diagnosis report, and removes the fault phase according to the diagnosis report so that the motor is converted to a fault-tolerant operation state.

[0043] This embodiment selects a 9-phase brushless DC motor drive system as an application example, and the stator winding of the motor is a single neutral point star connection symmetric structure, that is, the tail ends (or head ends) of the multiple windings of the motor are connected together to form a common point in the star connection, and the specific connection mode is as shown in Figure 1 The open-circuit fault diagnosis method of the multi-phase motor drive of the present application is based on the zero-state conservation equation of the multi-phase motor, and a fault and fault phase detection index is constructed according to the phase distribution diagram of the winding of the multi-phase motor; at the same time, the absolute value operation is introduced to further extract the potential fault information in the phase current to construct an upper and lower switch fault diagnosis index, so as to realize real-time open-circuit fault diagnosis. The method specifically includes current sampling, fault diagnosis index construction, fault and fault phase detection, and fault positioning. For the embodiment of the 9-phase brushless DC motor, the specific implementation steps of the method are as shown in Figure 3

[0044] Step (1): The motor phase current signal is obtained by current sensor sampling, and the zero-state conservation equation of the motor in the normal state and the fault state is constructed, that is, in the multi-phase motor, the sum of the current vectors of all phase windings is equal to zero at any time; and an initial fault index P of f phases is constructed f , wherein f represents the phase, that is, f = {A1, A2, A3, …, An}, A represents the general symbol of the phase, and n represents the number of phases of the motor.

[0045] For the normal state, the current of each phase is collected, and when the speed and torque reach the given value and are stable, the sampling time is the collection time, and the current value of each phase during normal operation is collected by the current sensor. For the fault state, the current of each phase is collected; the open-circuit fault of the inverter circuit IGBT in the nine-phase brushless DC motor drive system is divided into three categories, and the fault points are set as follows:

[0046] The first type is a phase-open fault, that is, the upper and lower switch tubes of each phase have an open-circuit fault;

[0047] The second type is an upper switch fault, that is, the fault caused by the open circuit of the upper switch tube of each phase;

[0048] The third type is a lower switch fault, that is, the fault caused by the open circuit of the lower switch of each phase.

[0049] The collection time is equal to the sampling time, and the current sensor collects the current value of each phase under the three fault states.

[0050] In the star-connected symmetric multi-phase motor drive system, according to the zero-state conservation equation, that is, the sum of the current vectors of all phase windings of the motor is equal to zero at any time, which is specifically shown in the following formula:

[0051]

[0052] In the formula: i k ​is the instantaneous value of the motor k-phase stator current signal (unit: ampere), N is the phase of the motor, n is the number of motor phases, and n>3.

[0053] When an open-circuit fault occurs in phase f, the current balance equation of the fault phase f can be written according to the above zero-state conservation constraints, as shown in the following formula:

[0054]

[0055] Where: f is the phase difference of the fault phase, and f = {A1, A2, A3, ..., An}; N-{f} represents the set containing all phase differences except the fth phase, and N∈{A1, A2, ..., An}; further, based on the f phase as the symmetry axis, the elements in N-{f} are divided into two sets Q f(n) and H f(n) ;Q f(n) It means that the fth phase is the symmetry axis, and the 2g phases adjacent in the counterclockwise direction are H f(n) Indicates the other n-2g-1 phases that are not in Q f(n) In the f-th phase, g represents a maximum integer not greater than n / 4; N-{f} represents a set including all phases except the f-th phase.

[0056] Specifically, for Q f(n) and H f(n) Describing mathematically, we get:

[0057]

[0058] Where: m represents the phase number corresponding to the fault phase, that is, f = Am, M is the phase number corresponding to the motor, and M∈{1,2,3,…,n}, n>3, mod is the modular operator symbol used to determine parity, ∧ is the logical AND operator symbol, and ∪ represents the union of sets.

[0059] Performing mathematical transformation on the above formula, we get the following:

[0060]

[0061] The initial fault index of phase f can be obtained as shown in the following formula:

[0062]

[0063] When the motor is in normal state, P f =0; when the motor f phase has an open circuit fault, P f =-1.

[0064] At the same time, it is important to note that each phase winding of a multi-phase BLDCM will be suspended during the commutation process. At this time, the current of the phase is zero, which also makes the above ratio instantaneously close to -1, which will reduce the robustness of the diagnosis method. To solve this problem, a judgment condition for the commutation duration should be set, that is, only when the ratio is close to -1 for multiple consecutive times is it judged as a fault. The specific method is to continuously sample P at a sampling frequency of f Signal, on this basis, the counter J is introduced to record P f The number of consecutive samplings is -1. f When P is -1, the counter J starts to accumulate; when P f When it is not equal to -1, the counter J is reset to zero. By setting the counter threshold C th To distinguish between normal commutation and fault status, if the counter J exceeds C th , it is a fault state; if it does not exceed, the system is considered to be in a normal state. The counting mechanism is used to avoid misjudgment caused by commutation, as shown in the following formula:

[0065]

[0066] Where: CT f is the counting mechanism fault detection variable, C th The threshold is the upper limit of the counter from the start to the end of healthy commutation. This threshold was obtained through multiple experiments. Furthermore, the extremely short commutation time and the subsequent combination of a hysteresis filter and a sliding average further effectively avoid misjudgments caused by commutation.

[0067] According to the above and Figure 2 As shown in the motor winding phase distribution, the initial fault diagnosis index after assuming an open-circuit fault in each phase is defined as follows:

[0068]

[0069]

[0070] Step (2): Filter the initial fault index and process the sliding average to obtain the final k-phase fault and fault phase detection index.

[0071] The initial fault diagnosis index P is analyzed by using hysteresis filtering algorithm. f Processing is performed to obtain the diagnostic index after filtering By introducing a minimum value ε, the initial fault diagnosis index normally outputs a constant value "1" in the range of -1-ε to -1+ε, and otherwise outputs a constant value "0", eliminating the misdiagnosis of the initial fault diagnosis index during normal operation. The specific hysteresis filtering algorithm is as follows:

[0072]

[0073] Where: ε is the hysteresis bandwidth, is the diagnostic index after filtering.

[0074] The filter module ensures that the current during normal operation of the motor will not affect the initial fault diagnosis index P f Generate interference, but the sudden addition of load to the motor will inevitably cause P f Enter the range of -1-ε and -1+ε, thereby avoiding loading misdiagnosis.

[0075] It is worth noting that the magnitude of load misdiagnosis is directly affected by the hysteresis bandwidth ε. As ε increases, the risk of false alarms also increases, while the speed of diagnosis also slightly increases. Considering the certain fluctuations in fault characteristics caused by measurement noise and other interference, in order to ensure that the fault indicator of the healthy phase has a low ripple, this width should be limited. Generally speaking, a 10% post-fault indicator value (i.e., ε = 0.1) usually provides a sufficient safety margin.

[0076] In order to improve the robustness of the algorithm, the fault diagnosis index after filtering is Introduce the sliding average algorithm to obtain the final fault index Its specific expression is shown as follows:

[0077]

[0078] Where: T δ is the period of the sliding average.

[0079] T δ =δT f

[0080] Where: T f is the fundamental period, which can be calculated by real-time speed; the parameter δ∈(0,1) defines the fundamental period part of the sliding average.

[0081] A lower δ results in a larger fault signature slope, speeding up fault signature extraction and ultimately diagnosis. However, this is often accompanied by higher ripple in the fault signature. Therefore, to achieve faster diagnosis while avoiding excessive ripple, δ is generally selected to be between 0.1 and 0.2.

[0082] Step (3): In order to display the fault information intuitively, the fault and fault phase detection flag SI is introduced f and fault and fault phase detection threshold T th ;like SI f =1, indicating that phase f has a fault; if SI f=0, indicating that phase f is healthy; from the above description, we can get the fault phase detection flag SI f Information:

[0083]

[0084] Where: T th The fault and fault phase detection threshold is selected by real-time calculation of the motor fault index in the healthy state. The maximum value obtained.

[0085] Step (4): In order to further extract the potential fault information contained in the phase current, the initial fault index P f Perform absolute value processing to obtain the fault switch detection index Its specific expression is shown in the following formula:

[0086]

[0087] Where: The absolute value operation is mainly used to preserve its polarity, so as to locate the fault more effectively.

[0088] Since the polarity is preserved, if the upper switch fails to open, If the lower switch tube has an open circuit fault, then

[0089] Step (5): In order to locate the faulty switch more effectively, two hysteresis filters are designed to construct the upper and lower switch fault diagnosis indices, which are r fu and r fl , u represents the upper switch tube, l represents the lower switch tube, and the specific filtering algorithm is shown in the following formula:

[0090]

[0091] Where: ε is the hysteresis bandwidth.

[0092] Step (6): In order to reduce misdiagnosis during loading operation, the upper and lower switch fault index r after filtering is fu and r fl Perform average calculation to obtain the final upper and lower switch fault monitoring index F fu and F fl , its specific expression is shown as follows:

[0093]

[0094] Where: r fu and r fl They represent the status monitoring indicators for the upper and lower switch tubes of phase f, ε u is the judgment threshold of the upper switch tube, εl is the judgment threshold of the lower switch tube, and <> is the average value calculation.

[0095] Due to the difference between the upper and lower switch tubes, the polarities of these two thresholds are opposite, that is, ε u ∈(0,1), ε l ∈(-1,0); however, through the average operation, the theoretical threshold value can be obtained, ε u =0.06,ε l =-0.06.

[0096] Step (7): According to the phase fault flag variable DI f =F fu +2*F fl , and combine the following rules to obtain the fault diagnosis results.

[0097] When DI f =0, it means phase f is healthy;

[0098] When DI f =1, it means that the switch tube on phase f has an open circuit fault;

[0099] When DI f =2, it means that the switch tube in phase f has an open circuit fault;

[0100] When DI f =3, it means that both the upper and lower switching tubes of phase f have open circuit faults, or the winding has a phase failure.

[0101] In order to more clearly and intuitively characterize the fault condition of each phase of the motor, a fault diagnosis table can be formed according to the rules described in step (7), as shown in the following table:

[0102] Table 1

[0103]

[0104]

[0105] The above description of the embodiments is intended to facilitate understanding and application of the present invention by those skilled in the art. It is apparent that those skilled in the art can readily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without requiring creative effort. Therefore, the present invention is not limited to the above embodiments. Any improvements or modifications made by those skilled in the art based on the disclosure of the present invention should fall within the scope of protection of the present invention.

Claims

1. A method for diagnosing open circuit faults in a motor drive system, characterized in that: The steps include: (1) Acquire the stator current signals of each phase of the multi-phase motor system and construct the initial fault index of each phase based on the zero-state conservation equation of the motor in normal and fault states; (2) continuously collecting and counting the initial fault index of each phase; (3) filtering and performing sliding average processing on the initial fault index to obtain a fault phase detection index for each phase; (4) Calculate and determine the fault phase detection flag of each phase based on the fault phase detection index and the count value and detect the fault phase accordingly; (5) For the fault phase, the initial fault index of the phase is processed by absolute value to obtain the fault switch detection index; (6) Passing the fault switch detection index through two hysteresis filters to obtain an upper switch tube fault diagnosis index and a lower switch tube fault diagnosis index; (7) performing an average calculation on the upper switch tube fault diagnosis index and the lower switch tube fault diagnosis index to obtain an upper switch tube state monitoring index and a lower switch tube state monitoring index; (8) Based on the upper switch tube state monitoring index and the lower switch tube state monitoring index, a positioning detection mark of the fault phase is calculated and determined to detect the fault position.

2. The method for diagnosing an open circuit fault in a motor drive system according to claim 1, wherein: The expression of the initial fault index in step (1) is as follows: Where: For any phase, f is the phase difference of the phase, P f is the initial fault index of the phase, i k is the instantaneous value of the k-phase stator current signal, Q f(n) It represents the 2g phases adjacent to each other in the counterclockwise direction with phase f as the symmetry axis, H f(n) Indicates the addition of f phase and Q f(n) For other phases except the first phase, g is the largest integer not greater than n / 4, and n is the number of motor phases.

3. The method for diagnosing an open circuit fault of a motor drive system according to claim 1, wherein: The specific implementation method of step (2) is: for any phase, the initial fault index of the phase is continuously collected at a fixed sampling frequency, and a counter is used to record the number of consecutive sampling times when the initial fault index is -1, that is, when the index value is -1, the counter accumulates 1; when the index value is not equal to -1, the counter is reset to 0.

4. The method for diagnosing an open circuit fault in a motor drive system according to claim 1, wherein: The specific implementation of step (4) is: for any phase, if And J>C th , then SI f =1, indicating that the phase has a fault; if and J≤C th , then SI f =0, indicating that the phase is healthy; is the fault phase detection index of the phase, J is the count value of the initial fault index of the phase, SI f is the fault phase detection flag of the phase, C th is the set counter threshold, T th is the set fault phase detection threshold.

5. The method for diagnosing an open circuit fault of a motor drive system according to claim 1, wherein: The transfer function expressions of the two hysteresis filters in step (6) are as follows: in: is the fault switch detection index of the fault phase, r fu and r fl are the upper switch fault diagnosis index and the lower switch fault diagnosis index of the fault phase, respectively, and ε is the hysteresis bandwidth of the hysteresis filter.

6. The method for diagnosing an open circuit fault in a motor drive system according to claim 1, wherein: The specific implementation of step (8) is as follows: first calculate the fault phase location detection mark DI according to the following formula f : DI f =F fu +2*F fl Among them: F fu and F fl They are the upper switch tube state monitoring index and the lower switch tube state monitoring index of the fault phase respectively; When DI f =1, it indicates that the upper switch tube of the fault phase has an open circuit fault; When DI f =2, it means that the lower switch tube of the fault phase has an open circuit fault; When DI f =3, it means that both the upper and lower switching tubes of the faulty phase have open circuit faults, or the winding has a phase failure.

7. The method for diagnosing an open circuit fault in a motor drive system according to claim 1, wherein: The stator winding of the multi-phase motor is a star-connected symmetrical structure with a single neutral point, that is, the tail ends of multiple windings of the motor are connected together to form a common point in the star connection.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: The processor is configured to execute the computer program to implement the method for diagnosing an open circuit fault of a motor drive system according to any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for diagnosing an open circuit fault of a motor drive system according to any one of claims 1 to 7 is implemented.

Images