A Five-Phase Motor Open-Circuit Fault Diagnosis Method

Abstract

The present invention discloses a method for diagnosing open - circuit faults of a five - phase motor, belonging to the field of on - line diagnosis of motors. In the model training stage, the method collects the time - series motor parameters of the motor under different working conditions through down - sampling technology to construct a training database; analyzes the relationship between the motor parameters and the winding current by using the Pearson correlation coefficient, and uses the BP neural network model to perform regression prediction on the motor state. In the on - line diagnosis stage, based on the trained digital - twin virtual motor, combined with the real - time collected electrical angle, torque and phase - voltage data of the physical motor for on - line training and current residual comparison to detect whether the motor has a fault. The fault location is accurately located by analyzing the difference value of the current residual and the diagnostic variable, and the fault type is judged as a winding open - circuit fault or a power - transistor open - circuit fault. The present invention does not require the accurate design parameters and fault data of the motor, and only needs the normal operation data to perform real - time and rapid diagnosis of the motor open - circuit fault and has strong robustness.

Description

Technical Field

[0001] The present invention relates to the field of on-line diagnosis of motors, and particularly to a method for diagnosing open-circuit faults of a five-phase motor. Background Art

[0002] In recent years, with the rapid rise of the low-altitude economy, the application of unmanned aerial vehicle (UAV) technology has become increasingly widespread in fields such as logistics transportation, agricultural monitoring, and environmental protection. As a core component of the low-altitude economy, the performance and reliability of UAVs directly affect the efficiency and safety of related industries. In the motor drive system of UAVs, on-line fault diagnosis of motors is particularly important. In particular, open-circuit faults in windings may cause UAVs to lose power or control, thereby triggering serious accidents. Currently, most existing on-line diagnosis methods rely on regular inspections and manual maintenance, unable to achieve real-time monitoring of the motor state, and lacking a timely emergency response mechanism. Therefore, there is an urgent need for an innovative on-line diagnosis technology for open-circuit faults in motor drive systems to improve the safety and reliability of UAVs, thereby providing guarantee for the sustainable development of the low-altitude economy.

[0003] Faults in motor drive systems usually include winding faults and power transistor faults. Existing on-line diagnosis methods for motors can be classified into signal processing methods, analytical model methods, and data-driven methods according to different selected characteristic values. However, these methods each have deficiencies in practical applications. Signal processing methods rely on extracting fault features from complex signals, but their diagnostic accuracy is low in high-noise environments, and at the same time, due to large computational amounts, the real-time performance is poor. Analytical model methods require an accurate mathematical model of the motor, but in practical applications, due to the inevitable differences between the mathematical model of the motor and the actual working conditions, their adaptability is poor. Especially in multi-parameter fault diagnosis, it is often difficult to handle complex fault situations, and the modeling process is complex, requiring high professional knowledge. Data-driven methods can establish a diagnostic database based on a large amount of fault data, but also require high computing resources and are often limited by the existing data acquisition and processing capabilities. Therefore, there is still room for improvement in the accuracy, adaptability, and real-time performance of existing on-line diagnosis methods, and further optimization is urgently needed to meet the diverse and highly real-time on-line diagnosis requirements of motor systems. Summary of the Invention

[0004] The purpose of the present invention is to overcome the above-mentioned defects in the background art and provide a method for diagnosing open-circuit faults of a five-phase motor, which does not require motor parameters during a fault, does not need to add additional hardware circuits, has a simple diagnostic algorithm, a fast diagnostic speed, a strong anti-misdiagnosis ability, and can quickly and accurately diagnose open-circuit faults in the windings and power transistors of the motor.

[0005] The present invention adopts the following technical solutions to solve the above technical problems:

[0006] A five-phase motor open-circuit fault diagnosis method, characterized in that the diagnosis method includes a model training stage and an online diagnosis stage, and the specific steps are as follows:

[0007] Step S1: In the model training stage, collect the time-series motor parameters of the five-phase motor under normal operation in different working conditions through downsampling technology, and construct a training database;

[0008] Step S2: In the model training stage, use the Pearson correlation coefficient to rank the importance of the collected five-phase motor parameters, and select the electrical angle θ(t), torque T e (t) and phase voltage u dc (t) at the same moment as the input data of the BP neural network regression model;

[0009] Step S3: In the model training stage, adopt an online training method, use the electrical angle, torque and phase voltage data of the normal operation in the training database, and combine the corresponding winding current data as the output, and obtain a preliminary BP neural network model through training;

[0010] Step S4: In the online diagnosis stage, save the BP neural network model and its parameters obtained in Step S3, and use it as a digital twin virtual motor; collect the electrical angle, torque and phase voltage of the physical motor at the same moment in real time through downsampling technology, and input these data into the digital twin virtual motor, and update the winding current i of the virtual motor through online training pn , where n represents the A, B, C, D, and E phases of the five-phase motor, and the parameters of the digital twin virtual motor are updated in real time;

[0011] Step S5: In the online diagnosis stage, compare the actual winding current i rn of the physical entity motor during operation with the winding current i pn of the digital twin virtual motor, and calculate the current residual i error_n ;

[0012] Step S6: In the online diagnosis stage, judge whether the motor has a winding open-circuit fault or a power transistor open-circuit fault by analyzing the difference value Δi of the current residual;

[0013] Step S7: In the online diagnosis stage, record the sampled current value i1 at 0.5 current fundamental periods after the fault moment and the sampled current value i2 at 0.75 current fundamental periods, perform summation processing on the two current values, and calculate the diagnosis variable K n for fault location.

[0014] Further, when collecting motor parameter data using downsampling technology to construct a database in step S1, the collected data is preprocessed, null values in the database are deleted, and obvious outliers in the database are removed using the Z-score method.

[0015] Further, in step S2, the Pearson correlation coefficient is calculated to evaluate the correlation between various motor parameters and winding current, which specifically includes the following steps:

[0016] Step S2-1: Randomly select a subset from the database as the sample set. Each motor parameter in the subset is used as the input set, and the corresponding winding current is used as the output set;

[0017] Step S2-2: Use the Pearson correlation coefficient formula to calculate the correlation coefficient between each motor parameter and winding current. The formula is as follows:

[0018]

[0019] where, Q i is the motor parameter sample, P i is the winding current sample, and are the average values of the corresponding samples respectively, and the value range of r Q,P is [-1, 1];

[0020] Step S2-3: Judge the influence degree of each motor parameter on winding current through the absolute value of the correlation coefficient. The higher the correlation, the more significant the influence of the parameter on winding current;

[0021] Step S2-4: Sort according to the calculated correlation coefficients, and screen out the key motor parameters that have a significant influence on winding current.

[0022] Further, the BP neural network model in step S2 is a neural network with a multi-hidden layer structure, which has three layers: the input layer, the hidden layer, and the output layer. Among them, the input ports of the input layer are 3. a is the number of data of each motor parameter in the database. The input layer ports are the electrical angle θ(t), the torque T e (t), and the phase voltage u dc (t); the output port of the output layer is 1, and the output is the winding current i of the digital twin virtual motor pn ; the number of neurons in its hidden layer is M, and the number of hidden layers is N. The specific values of N and M are adjusted according to the actual working conditions.

[0023] Further, in step S3, the BP neural network model is trained in an online training manner. The feature quantity x n input at time t is:

[0024] xn = [θ(t), T e (t), u dc (t)]

[0025] where x n is the dataset of electrical angle θ, torque T e , phase voltage u dc and the motor winding current at the previous moment in the database; θ(t) represents the electrical angle of the motor at time t in the database; T e (t) represents the electrical angle of the motor at time t, and u dc (t) represents the phase voltage of the motor at time t; the characteristic quantities are sequentially input into the BP neural network model according to the time sequence for model training.

[0026] Furthermore, in step S4, the trained BP neural network model is used as the digital twin virtual motor, and the downsampling technology is adopted to collect the electrical angle, torque and phase voltage data at the same moment during the operation of the physical entity motor in real time. The collected real-time data is input into the digital twin virtual motor model in the form of streaming data to obtain the digital twin virtual motor winding current i pn , and the historical data is processed by the sliding window method. The model only retains the data points of the most recent fundamental wave period at each time step as the real-time training set for online training.

[0027] Furthermore, in step S5, the current residual i error_n of the motor winding current at time t is:

[0028] i error_n (t) = i rn (t) - i pn (t)

[0029] where i rn (t) is the actual current of the physical entity motor winding at time t, and i pn (t) is the digital twin virtual motor winding current at time t. In step S6, the difference value Δi of the current residual of the motor winding is:

[0030]

[0031] where t n is t n-1 the next moment;

[0032] The judgment basis for online diagnosis is:

[0033]

[0034] where d1 is the set online diagnosis threshold.

[0035] Further, in step S7, the diagnostic variable K n has the following formula:

[0036] K n = i1 + i2

[0037] where i1 is the sampled current value at 0.5 fundamental current periods after the fault, and i2 is the sampled current value at 0.75 fundamental current periods after the fault;

[0038] The judgment basis for fault location is:

[0039]

[0040] Further, the five-phase permanent magnet synchronous motor drive system consists of a five-phase permanent magnet fault-tolerant motor and a set of five-phase half-bridge inverters. The power switch tubes T1, T2, T3, T4, T5, T6, T7, T8, T9, and T10 form the upper and lower arms of the five-phase half-bridge inverter. The inverter busbar end is supplied with DC power U dc The five-phase permanent magnet fault-tolerant motor adopts a 10-slot 8-pole fractional-slot double-layer concentrated winding structure. The electrical angle between two-phase slots is 144°, the windings have no overlap, the rotor adopts a surface-mounted permanent magnet structure, and the centrifugal height is 5 mm.

[0041] Compared with the prior art, the present invention adopts the above technical solutions and has the following beneficial effects:

[0042] (1) A five-phase motor open-circuit fault diagnosis method provided by the present invention does not need to collect various fault data of the motor, and only needs the normal operation data of the motor to diagnose the fault, greatly reducing the difficulty of diagnostic data collection.

[0043] (2) A five-phase motor open-circuit fault diagnosis method provided by the present invention can obtain accurate motor winding currents without prior knowledge of the precise parameters of the motor.

[0044] (3) A five-phase motor open-circuit fault diagnosis method provided by the present invention. The proposed downsampling technology can accurately collect different parameter data of the motor at the same moment, ensuring the time consistency of each parameter data and the effectiveness of the training data.

[0045] (4) A five-phase motor open-circuit fault diagnosis method provided by the present invention. The BP neural network model is trained using the healthy data during actual operation, so it can more accurately replicate the state of the target motion than the observer-based mathematical model.

[0046] (5) A five-phase motor open-circuit fault diagnosis method provided by the present invention uses a random forest model to extract the feature quantities required for training, without taking all the feature parameters as inputs, reducing the model training time and ensuring the accuracy of the model.

[0047] (6) A five-phase motor open-circuit fault diagnosis method provided by the present invention has an extremely fast diagnosis speed and a simple diagnosis algorithm, which is superior to general online diagnosis methods and does not require the formulation of complex thresholds.

[0048] (7) A five-phase motor open-circuit fault diagnosis method provided by the present invention uses the difference value of the current residual as the standard for judging whether a fault occurs, has strong robustness, and can well prevent misdiagnosis. Description of the Drawings

[0049] Figure 1 is the flow block diagram of a five-phase motor open-circuit fault diagnosis method of the present invention;

[0050] Figure 2 is the control block diagram of the five-phase motor of a five-phase motor open-circuit fault diagnosis method of the present invention;

[0051] Figure 3 is the digital twin virtual motor and the online diagnosis strategy flow chart of a five-phase motor open-circuit fault diagnosis method of the present invention. Detailed Embodiments

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings of the specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0053] Embodiment:

[0054] As Figure 1 shown, it is the flow block diagram of a five-phase motor open-circuit fault diagnosis method of the present invention, specifically including the following steps:

[0055] Step S1: In the model training stage, when using the down-sampling technology to collect the motor parameter data to construct the database, the collected data is preprocessed, the null values in the database are deleted, and the obvious outliers in the database are removed by using the Z-score method;

[0056] Step S2. In the model training stage, the Pearson correlation coefficient is used to rank the importance of the motor operation data in the database, and the most important parameters are selected to train the BP neural network model, reducing the training cost and time of the model. The random forest model consists of T decision trees, and the number of T is determined according to the actual working conditions, including the following steps:

[0057] Step S2-1. Randomly select a subset from the database as the sample set, where the motor parameters are used as the input set and the corresponding winding current is used as the output set;

[0058] Step S2-2. Use the Pearson correlation coefficient formula to calculate the correlation coefficient between each motor parameter and the winding current. The formula is as follows:

[0059]

[0060] where Q i is the motor parameter sample, P i is the winding current sample, Q and are the average values of the corresponding samples respectively, and the value range of r Q,P is [-1, 1].

[0061] Step S2-3. Judge the influence degree of each motor parameter on the winding current through the absolute value of the correlation coefficient. The higher the correlation, the more significant the influence of the parameter on the winding current;

[0062] Step S2-4. According to the calculated correlation coefficient ranking, screen out the key motor parameters that have a significant influence on the winding current.

[0063] Step S3. In the model training stage, use the electrical angle θ(t), torque T e (t), phase voltage u dc (t) and the previous moment motor winding current i(t - 1) data at the same moment in the database as the training set input. The BP neural network model is a multi-hidden layer neural network, including an input layer, a hidden layer and an output layer. Among them, the input port of the input layer is 4, a is the number of data of each motor parameter in the database, and the input layer ports are the electrical angle θ, torque T e and phase voltage u dc and the previous moment motor winding current i(t - 1); the output port of the output layer is 1, and the output is the digital twin virtual motor winding current i p ; the number of neurons in its hidden layer is M, and the specific value is adjusted according to the actual working conditions.

[0064] The BP neural network model is trained in an online training manner. The feature quantity x n input at time t is:

[0065] xn = [θ(t), T e (t), u dc (t), i(t - 1)]

[0066] where x n is the dataset of electrical angle θ, torque T e and phase voltage u dc in the database and the winding current of the physical entity motor at the previous moment; θ(t) represents the electrical angle of the motor at time t in the database; T e (t) represents the electrical angle of the motor at time t, u dc (t) represents the phase voltage of the motor at time t, and i(t - 1) represents the winding current of the motor at time t - 1; The characteristic quantities are input into the BP neural network model in chronological order for model training.

[0067] Step S4. In the online diagnosis stage, use the trained BP neural network model as the digital twin virtual motor, and adopt the downsampling technology to collect the electrical angle, torque, and phase voltage data of the physical entity motor at the same moment in real time. The collected real-time data is input into the digital twin virtual motor model in the form of streaming data to obtain the digital twin virtual motor winding current i p , and process the historical data through the sliding window method. The model only retains the data points of the most recent fundamental wave period as the real-time training set at each time step.

[0068] Step S5. In the online diagnosis stage, the current residual i error of the motor winding current at time t is:

[0069] i error (t) = i r (t) - i p (t)

[0070] where i r (t) is the actual current of the physical entity motor winding at time t, and i p (t) is the winding current of the digital twin virtual motor at time t.

[0071] Step S6. In the online diagnosis stage, use the difference value Δi of the current residual i error as the judgment criterion for whether the motor has winding open circuit faults and power transistor open circuit faults. The difference value Δi of the current residual of the motor winding is:

[0072]

[0073] where t n is the next moment after t n-1 ;

[0074] The judgment basis for online diagnosis is:

[0075]

[0076] where d is the set online diagnosis threshold.

[0077] Step S7: In the online diagnosis stage, record the sampled current value i1 at 0.5 current fundamental periods after the fault moment and the sampled current value i2 at 0.75 current fundamental periods after the fault moment, perform an addition process on the two current values, and calculate the diagnosis variable K n Perform fault location.

[0078] Diagnosis variable K n The formula for is:

[0079] K n = i1 + i2

[0080] where i1 is the sampled current value at 0.5 current fundamental periods after the fault, and i2 is the sampled current value at 0.75 current fundamental periods after the fault;

[0081] The judgment basis for fault location is:

[0082]

[0083] The five-phase permanent magnet synchronous motor drive system is composed of a five-phase permanent magnet fault-tolerant motor and a set of five-phase half-bridge inverters. The power switch tubes T1, T2, T3, T4, T5, T6, T7, T8, T9, and T10 constitute the upper and lower bridge arms of the five-phase half-bridge inverter, and the inverter bus terminal is U dc DC powered, the five-phase permanent magnet fault-tolerant motor adopts a 10-slot 8-pole fractional-slot double-layer concentrated winding structure, the electrical angle between two-phase slots is 144°, the windings have no overlap, the rotor adopts a surface-mounted permanent magnet structure, and the centrifugal height is 5 mm.

[0084] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for diagnosing open - circuit faults of a five - phase motor, characterized in that, The described diagnostic method includes a model training stage and an online diagnosis stage, and the specific steps are as follows: Step S1: In the model training stage, collect the time-series motor parameters of the five-phase motor during normal operation under different working conditions through downsampling technology, and construct a training database; Step S2: In the model training stage, use the Pearson correlation coefficient to rank the importance of the collected five-phase motor parameters, and select the electrical angle θ(t), torque T e (t), and phase voltage u dc (t) at the same moment as the input data of the BP neural network regression model; Step S3: In the model training stage, adopt an online training method. Use the electrical angle, torque, and phase voltage data during normal operation in the training database, combined with the corresponding winding current data as the output, and obtain a preliminary BP neural network model through training; Step S4: In the online diagnosis stage, save the BP neural network model and its parameters obtained in training in Step S3, and use it as the digital twin virtual motor; collect the electrical angle, torque, and phase voltage of the physical motor at the same time through downsampling technology, input these data into the digital twin virtual motor, and update the winding current i of the virtual motor through online training pn , where n represents the five-phase motor phases A, B, C, D, and E, and update the parameters of the digital twin virtual motor in real time; Step S5: In the online diagnosis stage, compare the actual winding current i of the physical entity motor during operation with the winding current i of the digital twin virtual motor, and calculate the current residual i rn ; pn ; error_n ; Step S6: In the online diagnosis stage, judge whether the motor has a winding open circuit fault or a power transistor open circuit fault by analyzing the difference value Δi of the current residual; Step S7: In the online diagnosis stage, record the sampled current value i1 at 0.5 current fundamental periods after the fault moment and the sampled current value i2 at 0.75 current fundamental periods, perform an addition process on the two current values, and calculate the diagnostic variable K n Perform fault location.

2. The open - circuit fault diagnosis method for a five - phase motor according to claim 1, wherein When using downsampling technology to collect motor parameter data to construct a database in Step S1, preprocess the collected data, delete the null values in the database, and use the Z-score method to remove the obvious outliers in the database.

3. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that, In Step S2, the correlation between each motor parameter and the winding current is evaluated by calculating the Pearson correlation coefficient, which specifically includes the following steps: Step S2-1: Randomly select a subset from the database as the sample set. The various motor parameters in the subset are used as the input set, and the corresponding winding current is used as the output set; Step S2-2: Use the Pearson correlation coefficient formula to calculate the correlation coefficients between each motor parameter and the winding current. The formula is as follows: Among them, Q i is the motor parameter sample, P i is the winding current sample, and are the average values of the corresponding samples respectively, and the value range of r Q,P is [-1, 1]; Step S2-3: Judge the influence degree of each motor parameter on the winding current through the absolute value of the correlation coefficient. The higher the correlation, the more significant the influence of the parameter on the winding current; Step S2-4: Sort according to the calculated correlation coefficients, and screen out the key motor parameters that have a significant influence on the winding current.

4. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that The BP neural network model in step S2 is a neural network with a multi-hidden layer structure, which has three layers: an input layer, a hidden layer, and an output layer. Among them, the input ports of the input layer are 3, a is the number of data of each motor parameter in the database, and the input layer ports are the electrical angle θ(t), the torque T e (t), and the phase voltage u dc (t); the output port of the output layer is 1, and the output is the current i of the digital twin virtual motor winding pn ; the number of neurons in its hidden layer is M, and the number of hidden layers is N. The specific values of N and M are adjusted according to the actual working conditions.

5. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that, In the step S3, the BP neural network model is trained by an online training method, and the feature quantity x input at the moment t n is as follows: x n = [θ(t), T e (t), u dc (t)] where x n is the data set of electrical angle θ, torque T e , phase voltage u dc and the motor winding current at the previous moment in the database; θ(t) represents the electrical angle of the motor at time t in the database; T e (t) represents the electrical angle of the motor at time t, u dc (t) represents the phase voltage of the motor at time t; the characteristic quantities are input into the BP neural network model in chronological order for model training.

6. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that In step S4, the trained BP neural network model is used as the digital twin virtual motor, and the downsampling technology is adopted to collect the electrical angle, torque, and phase voltage data of the physical entity motor at the same moment in real time. The collected real-time data is input into the digital twin virtual motor model in the form of streaming data to obtain the digital twin virtual motor winding current i pn , and the historical data is processed by the sliding window method. The model only retains the data points of the most recent fundamental wave period at each time step as the real-time training set for online training.

7. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that The current residual i of the motor winding current at time t in the step S5 error_n is as follows: i error_n i(t) = i rn i(t) - i pn i(t) where i rn (t) is the actual current of the physical entity motor winding at time t, and i pn (t) is the current of the digital twin virtual motor winding at time t. The difference value Δi of the current residual of the motor winding in step S6 is as follows: where t n is t n-1 the later moment; The judgment basis for online diagnosis is: where d1 is the set online diagnosis threshold.

8. A five-phase motor open-circuit fault diagnosis method according to claim 1, characterized in that The diagnostic variable K in the step S7 n has the formula: K n = i1 + i2 where i1 is the sampled current value at 0.5 current fundamental periods after the fault, and i2 is the sampled current value at 0.75 current fundamental periods after the fault; The judgment basis for fault location is:

9. A method for diagnosing open - circuit faults of a five - phase motor according to claim 1, characterized in that The five-phase permanent magnet synchronous motor drive system consists of a five-phase permanent magnet fault-tolerant motor and a set of five-phase half-bridge inverters. The power switching transistors T1, T2, T3, T4, T5, T6, T7, T8, T9, and T10 form the upper and lower arms of the five-phase half-bridge inverters. The inverter bus terminal is powered by a DC power supply U dc The five-phase permanent magnet fault-tolerant motor adopts a 10-slot 8-pole fractional-slot double-layer concentrated winding structure. The electrical angle between two-phase slots is 144°. The windings have no overlap. The rotor adopts a surface-mounted permanent magnet structure with a centrifugal height of 5 mm.

Images