Permanent magnet motor fault diagnosis method and system based on multi-electrical signal fusion

Abstract

The invention discloses a permanent magnet motor fault diagnosis method based on fusion of multiple electrical signals, and belongs to the technical field of motor fault diagnosis. According to the method, permanent magnet motor fault diagnosis is carried out by utilizing a method of fusing stator current signals which can be acquired in real time and tachometer signal features, the dependence on an additional sensor is solved, and the detection cost is reduced by utilizing a current sensor and an encoder in a driving system; meanwhile, as stator current and rotating speed signals can be collected in real time, online fault detection can be achieved, the real-time performance of diagnosis is improved, the two signals are not prone to being affected by environmental noise, the extraction precision of fault features in the signals is greatly improved, and accurate fault diagnosis is facilitated; besides, the accurate phase information of the rotor position is extracted from the stator current, so that the method can be suitable for fault diagnosis under the unsteady working condition of the motor, and the method is more competitive in practical application.

Description

Technical Field

[0001] The present invention belongs to the technical field of motor fault diagnosis, and in particular relates to a permanent magnet motor fault diagnosis method and system based on multi-electrical signal fusion. Background Art

[0002] The permanent magnet motor in the traction drive system needs to run at high speed for a long time and with heavy loads, and often has problems such as improper lubrication and untimely maintenance, which can cause the motor to easily fail in various forms. Among them, bearing failure is one of the most likely failure forms, which can directly cause the entire traction drive system to shut down.

[0003] At present, the commonly used fault diagnosis method is to collect various signals when the motor is running, analyze the signals, extract the information related to the fault, and then determine whether the fault has occurred and the type of fault. At present, the signals commonly used for fault diagnosis include: vibration signals, temperature signals, noise signals, etc. These signals often require additional sensors to be installed on the motor, which not only increases the cost of the detection system, but also requires the motor to be disconnected from the traction system, which makes it impossible to achieve online real-time detection. In addition, the signal acquisition process has high requirements for the acquisition environment and operation, otherwise it will be easy to cause noise interference in the collected signal, making it difficult to extract fault characteristics.

[0004] At present, the commonly used methods for permanent magnet motor fault diagnosis can be divided into two categories: data-driven intelligent classification methods and feature analysis methods based on signal processing.

[0005] The data-driven intelligent classification method mainly uses a large amount of motor signal data to train the constructed network model, extracts hidden features related to the fault in the signal, and uses these extracted features to make the algorithm converge and obtain a network model corresponding to different states of the motor. In actual diagnosis, the experimental data of the motor is input into the model, and then the automatic identification and classification of the fault is realized. However, this type of method requires a large amount of data to train the model, which is difficult to achieve in the actual diagnosis scenario of rail transit. At the same time, the training and testing process requires a long calculation time, which results in only offline fault detection and lack of real-time performance.

[0006] The feature analysis method based on signal processing is to extract the frequency components related to the fault in the signal through filtering, denoising, and transformation, and then diagnose the fault type by comparing it with the theoretical characteristics. However, the current signal processing method has a complex calculation process, high requirements for diagnostic personnel, and requires some algorithm parameters to be set in advance, which increases the difficulty of signal processing and reduces the efficiency of fault diagnosis. At the same time, the biggest disadvantage is that it cannot accurately extract fault harmonics.

[0007] In addition, the current fault diagnosis method has strict requirements on the motor operating conditions, and often requires the motor to rotate at a steady-state speed to collect steady-state fault signals. However, in actual engineering applications, the motor in the drive system often needs to accelerate and decelerate repeatedly, which will cause the collected signal to be a non-steady-state signal, which increases the difficulty of signal processing and feature extraction, and also reduces the accuracy of fault diagnosis. The current fault diagnosis method under variable speed conditions mainly uses an algorithm that combines angular domain resampling and order analysis. The core of this method is to obtain an accurate phase angle signal. The current commonly used solution is to install an additional tachometer to obtain rotor rotation information to obtain the phase angle, but this method requires the installation of additional sensors, which undoubtedly increases the cost of detection. Summary of the invention

[0008] The purpose of an embodiment of the present invention is to provide a permanent magnet motor fault diagnosis method based on multi-electrical signal fusion, which uses a method of fusing stator current signals and tachometer signal features that can be obtained in real time to perform permanent magnet motor fault diagnosis, thereby solving the dependence on additional sensors and reducing the detection cost by using current sensors and encoders in the drive system; at the same time, since the stator current and speed signals can be collected in real time, online fault detection can be achieved, which improves the real-time nature of diagnosis; at the same time, these two signals are not easily affected by environmental noise, which will greatly improve the extraction accuracy of fault features in the signal and is conducive to accurate diagnosis of faults; in addition, the present invention extracts precise rotor position phase information from the stator current, which can be applied to fault diagnosis under non-steady-state conditions of the motor, making it more competitive in practical applications, thereby solving at least one technical problem involved in the background technology.

[0009] In order to solve the above-mentioned technical problems, the present invention is achieved as follows:

[0010] The embodiment of the present invention provides a permanent magnet motor fault diagnosis method based on multi-electrical signal fusion, comprising the following steps:

[0011] Step S1, calculating the fault characteristic frequency in the speed signal when the bearing fails;

[0012] Step S2, collecting the stator current signal and the speed signal in the permanent magnet motor control system;

[0013] Step S3, using an empirical Fourier decomposition algorithm to decompose the collected stator current signal and speed signal, and extracting fault harmonics and current fundamental waves from the decomposed components;

[0014] Step S4, determining whether the motor is running at a constant speed, if so, executing step S5; if not, executing step S6;

[0015] Step S5, performing envelope spectrum analysis on the fault harmonics to obtain a frequency spectrum, and obtaining frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination;

[0016] Step S6, using the extracted current fundamental wave to obtain rotor phase information to obtain a continuous angle increment signal of the rotor;

[0017] Step S7, using the obtained continuous angle increment signal to perform angle domain resampling on the extracted fault harmonics, so as to convert the time domain signal into an angle domain signal;

[0018] Step S8, performing envelope spectrum analysis on the resampled angular domain signal, and determining whether a fault occurs and the type of the fault by comparing the fault information in the spectrum with the theoretical value.

[0019] Optionally, in step S1, the fault characteristic frequency includes an outer ring fault characteristic frequency and an inner ring fault characteristic frequency, which are expressed by the following formula:

[0020]

[0021] Among them, f ORF 、f IRF They are the characteristic frequency of outer ring fault and inner ring fault respectively; FCO ORF 、FCO IRF分 They are the characteristic frequency coefficient of outer ring fault and the characteristic frequency coefficient of inner ring fault respectively; N B is the number of bearing balls; D b D is the ball diameter; p is the ball diameter; α is the bearing contact angle; f r is the mechanical rotation frequency of the motor.

[0022] Optionally, in step S2, the stator current signal and the speed signal are respectively obtained by a current sensor and an incremental encoder in a permanent magnet motor control system.

[0023] Optionally, in step S6, the extracted current fundamental wave i(t) is used to obtain the rotor phase information to obtain a continuous angle increment signal C of the rotor. φ (t), including:

[0024] The current fundamental wave i(t) consists of two parts: amplitude component and phase component, which can be expressed as follows:

[0025]

[0026] Among them, f i (t) is the amplitude part; is the phase part, where j is the phase unit, It is obtained by Hilbert transform, as shown below:

[0027]

[0028] H represents the Hilbert transform, τ is the integral variable, and the instantaneous phase φ(t) of the current fundamental wave i(t) is calculated by the following formula:

[0029]

[0030] The continuous angle increment signal C of the rotor is obtained by processing φ(t) through the accumulation function unwrap() in Matlab. φ (t).

[0031] Optionally, in step S7, the angular domain resampling is expressed by the following formula:

[0032] Rf(θ)=resample(f(t),C φ (t)) (17)

[0033] Wherein, resample() is an existing function in MATLAB, and f(t) is the fault harmonic.

[0034] The present invention also provides a permanent magnet motor fault diagnosis system based on multi-electrical signal fusion for implementing the method described above, comprising:

[0035] A fault characteristic frequency calculation module, which is used to calculate the fault characteristic frequency in the speed signal when the bearing fails;

[0036] A stator current signal and speed signal acquisition module, which is used to acquire the stator current signal and speed signal in the permanent magnet motor control system;

[0037] A fault harmonic and current fundamental wave extraction module is used to decompose the collected stator current signal and speed signal using an empirical Fourier decomposition algorithm, and extract the fault harmonic and current fundamental wave from the decomposed components;

[0038] A working condition judgment module is used to judge whether the motor is running under a constant speed condition;

[0039] A fault diagnosis and type discrimination module is used to perform envelope spectrum analysis on fault harmonics to obtain a frequency spectrum, and obtain frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination;

[0040] An angle increment signal acquisition module, which is used to obtain rotor phase information using the extracted current fundamental wave to obtain a continuous angle increment signal of the rotor;

[0041] An angular domain signal conversion module, which is used to perform angular domain resampling on the extracted fault harmonics using the obtained continuous angular increment signal, so as to convert the time domain signal into an angular domain signal;

[0042] The fault determination and type discrimination module is used to perform envelope spectrum analysis on the resampled angular domain signal, and to determine whether a fault has occurred and to discriminate the fault type by comparing the fault information in the spectrum with the theoretical value.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. Compared with the conventional existing technology based on vibration signals, this technical solution utilizes the stator current and speed signals that can be measured in real time in the permanent magnet motor drive system for fault diagnosis, which can avoid the installation of additional sensors to reduce the cost of detection. At the same time, since these two signals can be obtained in real time, this technical solution can realize online detection. In addition, compared with vibration signals, stator current and speed signals are not easily affected by environmental noise, have a higher signal-to-noise ratio, and are more conducive to the extraction of fault harmonics.

[0045] 2. Compared with the existing technology based on data-driven intelligent diagnosis methods, this technical solution uses signal processing methods to analyze the collected signals, and then extracts the fault-related information in the signals to diagnose the bearing fault and distinguish the type. This technical solution does not require a large amount of data, does not require operators to perform complex work, and has relatively low theoretical requirements for operators. It does not require professional personnel to use it, so it is more suitable for practical engineering applications.

[0046] 3. Compared with the existing technology based on signal processing methods, the signal processing algorithm used in this technical solution is simple and fast, does not require the algorithm parameters to be set in advance, and can quickly and accurately extract the harmonic components related to the fault in the signal. At the same time, the rotor phase information obtained through the stator current is more accurate, which is more conducive to fault diagnosis under non-steady-state speed conditions.

[0047] 4. Compared with the existing fault diagnosis technology, this technical solution does not have strict requirements on the working conditions of the permanent magnet motor. The existing technology usually requires the motor speed to be constant in order to extract the fault characteristic frequency. However, in actual operating conditions, the motor is often in the process of acceleration and deceleration, which causes the motor speed to be non-stable, and thus also makes the collected signal a non-steady-state signal. This technical invention resamples the rotor position angle signal obtained by the stator current using the time domain non-steady-state signal to obtain a stable angular domain signal, thereby realizing fault diagnosis under different speed conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0049] Figure 1 A flowchart of a permanent magnet motor fault diagnosis method based on multi-electrical signal fusion provided by an embodiment of the present invention;

[0050] Figure 2 A bearing model diagram provided by an embodiment of the present invention;

[0051] Figure 3 A block diagram of motor control and signal acquisition provided by an embodiment of the present invention;

[0052] Figure 4 A structural block diagram of a permanent magnet motor fault diagnosis system based on multi-electrical signal fusion provided by an embodiment of the present invention;

[0053] Figure 5 An experimental platform diagram provided for an embodiment of the present invention;

[0054] Figure 6 The speed waveform diagrams under different bearings provided by the embodiment of the present invention, wherein: Figure 6 (a) is the speed waveform of a normal bearing; (b) is the speed waveform of a bearing with an outer ring fault;

[0055] Figure 7 The speed frequency diagram under different bearings provided by the embodiment of the present invention, wherein: Figure 7 (a) is the speed frequency diagram of a healthy bearing; 7(b) is the speed frequency diagram of a bearing with an outer ring fault;

[0056] Figure 8 The current signal comparison result diagram under different bearing states provided by the embodiment of the present invention, wherein: Figure 8 (a) is the time domain waveform; 8(b) is the spectrum diagram;

[0057] Fig. 9 This is a feature extraction result diagram under the outer ring fault bearing state provided by an embodiment of the present invention, wherein: Fig. 9 (a) is the fault harmonic diagram in the speed signal; 9(b) is the stator current fundamental wave diagram;

[0058] Fig.10 The extracted rotor position angle signal diagram provided by the embodiment of the present invention, wherein: Fig.10 (a) is the original image; 10(b) is the enlarged image;

[0059] Fig.11 The outer race fault diagnosis result diagram provided by the embodiment of the present invention, wherein: Fig.11 (a) is the spectrum diagram; 11(b) is the order diagram;

[0060] Fig.12 This is a diagram of inner race fault diagnosis results provided by an embodiment of the present invention, wherein: Fig.12 (a) is the spectrum diagram; 12(b) is the order diagram;

[0061] Fig.13 The variable speed speed waveform diagram provided by the embodiment of the present invention, wherein: Fig.13 (a) is the acceleration graph; 13(b) is the deceleration graph;

[0062] Fig.14 This is a diagram of the diagnostic results under acceleration conditions provided by an embodiment of the present invention, wherein: Fig.14 (a) is the outer ring fault diagram; 14(b) is the inner ring fault diagram;

[0063] Fig.15 This is a diagram of the diagnostic results under the deceleration condition provided by an embodiment of the present invention, wherein: Fig.15 (a) is the outer ring fault diagram; 15(b) is the inner ring fault diagram. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The terms "first", "second", etc. in the specification and claims of the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0066] See also Figure 1 As shown, an embodiment of the present invention provides a permanent magnet motor fault diagnosis method based on multi-electrical signal fusion, comprising the following steps:

[0067] Step S1, calculating the fault characteristic frequency in the speed signal when the bearing fails;

[0068] Step S2, collecting the stator current signal and the speed signal in the permanent magnet motor control system;

[0069] Step S3, using an empirical Fourier decomposition algorithm to decompose the collected stator current signal and speed signal, and extracting fault harmonics and current fundamental waves from the decomposed components;

[0070] Step S4, determining whether the motor is running at a constant speed, if so, executing step S5; if not, executing step S6;

[0071] Step S5, performing envelope spectrum analysis on the fault harmonics to obtain a frequency spectrum, and obtaining frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination;

[0072] Step S6, using the extracted current fundamental wave to obtain rotor phase information to obtain a continuous angle increment signal of the rotor;

[0073] Step S7, using the obtained continuous angle increment signal to perform angle domain resampling on the extracted fault harmonics, so as to convert the time domain signal into an angle domain signal;

[0074] Step S8, performing envelope spectrum analysis on the resampled angular domain signal, and determining whether a fault occurs and the type of the fault by comparing the fault information in the spectrum with the theoretical value.

[0075] In step S1, the fault characteristic frequency includes the outer ring fault characteristic frequency and the inner ring fault characteristic frequency, which is expressed by the following formula:

[0076]

[0077] Among them, f ORF 、f IRF They are the characteristic frequency of outer ring fault and inner ring fault respectively; FCO ORF 、FCO IRF分 They are the outer ring fault characteristic frequency coefficient and the inner ring fault characteristic frequency coefficient, also known as the fault order, which is related to the bearing size parameters. Figure 2 As shown, N B is the number of bearing balls; D b D is the ball diameter; p is the ball diameter; α is the bearing contact angle, which is generally 0 for rolling bearings; f r is the motor mechanical rotation frequency, which can be calculated from the motor speed n. The calculation formula is:

[0078]

[0079] The vibration of the motor will cause torque fluctuations, which will in turn cause harmonic components caused by faults in the motor speed signal. Therefore, fault diagnosis is performed by collecting the motor speed signal. The formula for speed harmonics caused by the fault is derived as follows:

[0080] When a bearing fails, the load torque can be seen as a function that changes with time, which mainly includes two parts: the load torque constant term T const and with the fault frequency f c Varying additional torque T add , that is, the load torque is expressed as:

[0081] T load (t) = T const +T add (t) (3)

[0082] Among them, T add (t) Without considering the high-order frequency components, that is, only considering the influence of the torque fundamental frequency caused by the bearing fault, the additional torque T add (t) can be approximated as c The cosine function at is expressed as:

[0083] T add (t) = T c cos(ω c t) (4)

[0084] Where: T c is the magnitude of the additional torque caused by the bearing fault, ω c =2πf c ;

[0085] Combining formula (3) and (4), the motor load formula can be obtained as follows:

[0086] T load (t) = T const +T c cos(ω c t) (5)

[0087] According to the torque equation of the motor:

[0088]

[0089] Where: T e is the electromagnetic torque generated by the motor; ω r is the mechanical angular velocity of the motor rotor; J is the total mechanical moment of inertia of the motor and load; B is the damping coefficient;

[0090] Ignoring the influence of damping, equation (6) can be transformed into:

[0091]

[0092] In the steady state, assuming that the electromagnetic torque T generated by the motor e is a steady-state value. Combining formulas (5) and (6), the constant term T of the load torque is const and electromagnetic torque T e Equivalent to a constant term C, formula (7) is transformed into:

[0093]

[0094] From formula (8), we can see that after the bearing failure occurs, the impact on the motor speed is at the constant speed value ω r0 A time-varying sine term is added to it.

[0095] According to the rotor angle θ r is the mechanical speed ω r The integral relationship gives the rotor angle θ r The expression is:

[0096]

[0097] Conclusion: The bearing fault causes the change of motor load torque, which in turn causes the motor rotor position and speed to fluctuate. Since the motor torque is linearly related to the speed, the fault characteristics of the bearing fault in the motor rotor position and speed signals are the same as the fault frequency in the vibration signal.

[0098] It should be further explained that, considering the impact of bearing fault on stator current, it is proved that even in the event of fault, the stator current is still dominated by the fundamental harmonics of the current, which does not affect the accuracy of the extracted rotor phase information. The derivation process is as follows:

[0099] When a motor fails, the stator current is modulated by the fault harmonics. At this time, the expression of the stator current is as follows:

[0100] i F (t) = i H (t)[1+βcos(ω c t)] (10)

[0101] Among them, i F (t) is the stator current under fault condition, i H (t) is the stator current in a healthy state, β is the fault modulation coefficient, ω c is the fault angular frequency.

[0102] Since the fault harmonics are divided into harmonics of different orders, formula (10) is expressed in detail as follows:

[0103]

[0104] Among them, β k is the modulation coefficient of each order fault harmonic, and k is a positive integer.

[0105] Stator current i in healthy state H (t) is expressed as:

[0106] i H (t) = I m cos(ω e t)=I m cos(2πf e t) (12)

[0107] Among them, I m is the current amplitude, ω e is the current angular frequency, f e is the current frequency.

[0108] Where, the current angular frequency ω e and the current frequency f e and the speed angular frequency ω r and speed frequency f r The relationship can be expressed as:

[0109]

[0110] Among them, p n is the number of motor pole pairs. From the formula, we can see that the instantaneous frequency of the stator current is faster than the frequency of the speed, so the phase information is richer and the extracted rotor phase angle is more accurate.

[0111] Combining equations (11) and (12), we can get the detailed stator current expression under fault conditions:

[0112]

[0113] Since the value of the fault modulation coefficient is very small, the stator current is still dominated by its fundamental frequency component. Therefore, it is feasible to extract accurate rotor phase information from the measured stator current.

[0114] In step S2, combined Figure 3 As shown, the stator current signal and the speed signal are obtained through the current sensor and the incremental encoder in the permanent magnet motor control system respectively. These two signals are also used as feedback signals for the motor closed-loop control. Therefore, these two signals need to be acquired in real time, which provides conditions for low-cost, online and real-time fault diagnosis.

[0115] It should be noted that Figure 3 In, ω ref is the reference angular velocity; ω e is the feedback angular velocity; iref is the reference quadrature-axis current; i q is the feedback quadrature axis current; i ref is the reference direct axis current; i d is the feedback direct axis current; u q * is the reference quadrature axis voltage; u d * is the reference direct-axis voltage; i a 、i b 、i c is the three-phase current; θ is the motor rotor position angle; SVPWM is space vector modulation; PMSM is permanent magnet synchronous motor; abc / dq is park transformation; dq / abc is inverse park transformation; LPF is low-pass filter.

[0116] Figure 3 In the present invention, the motor control method adopts a dual closed-loop vector control system, and the control object is a permanent magnet synchronous motor. The closed-loop control is divided into speed loop control and current loop control. In order to realize the closed-loop control of the motor speed, a 2500-line incremental encoder is installed on the motor rotor shaft, which can not only make the PMSM run stably at the set speed, but also obtain the motor rotor position information θ. By differential processing the rotor position information, the speed signal of the motor can be obtained when it is running. Therefore, the detection system does not need to install additional sensors, which reduces the detection cost.

[0117] In step S6, the extracted current fundamental wave i(t) is used to obtain the rotor phase information to obtain the continuous angle increment signal C of the rotor. φ (t), including:

[0118] The current fundamental wave i(t) consists of two parts: amplitude component and phase component, which can be expressed as follows:

[0119]

[0120] Among them, f i (t) is the amplitude part; is the phase part, where j is the phase unit, It is obtained by Hilbert transform, as shown below:

[0121]

[0122] H represents the Hilbert transform, τ is the integral variable, and the instantaneous phase φ(t) of the current fundamental wave i(t) is calculated by the following formula:

[0123]

[0124] Since the amplitude of the phase angle φ(t) is from -π to π, a continuously increasing angle signal is required for angular domain resampling. The φ(t) is processed by the accumulation function unwrap() in Matlab to obtain the continuous angle increment signal C of the rotor. φ (t).

[0125] In step S7, the angular domain resampling is expressed by the following formula:

[0126] Rf(θ)=resample(f(t),C φ (t)) (17)

[0127] Wherein, resample() is an existing function in MATLAB, and f(t) is the fault harmonic.

[0128] Recombination Figure 4 As shown, the present invention also provides a permanent magnet motor fault diagnosis system based on multi-electrical signal fusion for implementing the method described, including a fault characteristic frequency calculation module 10, a stator current signal and speed signal acquisition module 20, a fault harmonic and current fundamental wave extraction module 30, a working condition judgment module 40, a fault diagnosis and type discrimination module 50, an angle increment signal acquisition module 60, an angle domain signal conversion module 70 and a fault determination and type discrimination module 80.

[0129] The fault characteristic frequency calculation module 10 is used to calculate the fault characteristic frequency in the rotation speed signal when a bearing fails.

[0130] The stator current signal and speed signal acquisition module 20 is used to acquire the stator current signal and speed signal in the permanent magnet motor control system.

[0131] The fault harmonic and current fundamental wave extraction module 30 is used to decompose the collected stator current signal and speed signal by using an empirical Fourier decomposition algorithm, and extract the fault harmonic and current fundamental wave from the decomposed components.

[0132] The operating condition determination module 40 is used to determine whether the motor is operating under a constant speed condition.

[0133] The fault diagnosis and type discrimination module 50 is used to perform envelope spectrum analysis on the fault harmonics to obtain a frequency spectrum, and obtain frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination;

[0134] The angle increment signal acquisition module 60 is used to obtain rotor phase information using the extracted current fundamental wave to obtain a continuous angle increment signal of the rotor.

[0135] The angle domain signal conversion module 70 is used to perform angle domain resampling on the extracted fault harmonics using the obtained continuous angle increment signal, so as to convert the time domain signal into an angle domain signal.

[0136] The fault determination and type discrimination module 80 is used to perform envelope spectrum analysis on the resampled angular domain signal, and to determine whether a fault occurs and to discriminate the fault type by comparing the fault information in the spectrum with the theoretical value.

[0137] The permanent magnet motor fault diagnosis method and system based on multi-electrical signal fusion provided by the present invention are described in detail below with specific embodiments.

[0138] First, build an experimental platform to collect experimental data, see Figure 5 As shown, the experimental platform includes a faulty motor 1, a load motor 2, a circuit board 3, a controller 4, a computer 5, an encoder 6 and a vibration sensor 7.

[0139] The model of the faulty bearing is SKF6305, and its size parameters as well as the characteristic frequencies and coefficients of different faults are shown in Table 1.

[0140] Table 1 Bearing size parameters

[0141]

[0142]

[0143] First, the encoder signals of the healthy bearing and the outer ring faulty bearing at a speed of 1000r / min are collected and converted into speed signals, whose waveforms are as follows: Figure 6 As shown. Figure 6 (a) and Figure 6 (b) It can be seen that when a bearing failure occurs, the speed fluctuation of the permanent magnet motor increases significantly.

[0144] Secondly, in order to verify that the speed signal can obtain information related to the fault, the spectrum analysis of the two groups of speed signals is performed. The results are as follows: Figure 7 As shown. Figure 7 (a) and Figure 7 (b) It can be seen that in the speed signal spectrum of both the healthy bearing and the faulty bearing, at 1000 r / min, the speed base frequency and the frequency multiplication component nf r (f r =16.67Hz) is the dominant component, resulting in fault characteristics and its multiple frequency nf ORF (f ORF =43.24Hz) is submerged, making it impossible to diagnose and identify the bearing type.

[0145] In order to verify that the fault will not produce excessive harmonic effects on the stator current, the stator current waveform and spectrum of the healthy bearing and the outer ring fault bearing are compared, such as Figure 8 shown.

[0146] Depend on Figure 8 (a) and Figure 8 (b) It can be seen that even when a fault occurs, the fundamental frequency component is still dominant in the stator current. Therefore, it is feasible to obtain the instantaneous phase of the rotor by extracting the fundamental frequency component.

[0147] In order to extract the fault characteristic harmonics and fundamental components of the stator current in the speed signal, the speed signal and stator current under the outer ring fault are empirically decomposed by Fourier transform. The fault harmonics and current fundamental components are as follows: Fig. 9 shown.

[0148] Depend on Fig. 9 (a) It can be seen that after the empirical Fourier decomposition, the fault harmonics extracted from the speed signal eliminate the influence of the speed fundamental wave and only retain the harmonic components caused by the fault. Fig. 9 (b) It can be seen that the extracted current fundamental wave is more sinusoidal than the original current waveform under fault, so it is beneficial to extract the accurate rotor phase from it to improve the accuracy of angular domain resampling. In order to verify that the rotor phase can be extracted from the stator current, the method proposed in the present invention is used to process the extracted current fundamental wave, and the angle obtained is as follows: Fig.10 shown.

[0149] pass Fig.10 (a) and Fig.10 (b) It can be seen that the rotor angle obtained from the stator current fundamental wave by the technology proposed in the present invention is almost the same as the reference angle obtained by the encoder, which verifies the effectiveness of the proposed method.

[0150] Therefore, the extracted fault feature components are resampled using the obtained angle signal. The results before and after resampling are as follows: Fig.11 shown.

[0151] Depend on Fig.11 As shown in (a), after the speed signal is decomposed and the fault characteristic harmonics are extracted, the fault characteristic frequency and its multiple frequency nf in the spectrum ORF (f ORF =43.24Hz) becomes obvious and is almost consistent with the theoretical value of 43.30Hz, so it can be determined that an outer race fault has occurred. After resampling the fault harmonics using the extracted rotor angle signal, the analysis results are as follows: Fig.11 As shown in (b), the main components and double frequencies in the order spectrum are nFCO ORF (FCO ORF=2.60), which also corresponds to the characteristic order of the outer race fault. Therefore, the effectiveness of the proposed technology can be proved.

[0152] The same processing is performed on the speed signal under inner race fault, and the results are as follows: Fig.12 shown.

[0153] Fig.12 In (a), the fault characteristics and the frequency multiplication nf IRF (f IRF =73.20Hz) is almost consistent with the theoretical value and becomes very obvious, which is conducive to fault diagnosis and fault type identification. At the same time, in 12(b), the fault characteristics are converted into orders, and the inner ring order is nFCO IRF (FCO IRF =4.40) is the main component, which corresponds to the theoretical order of inner race fault.

[0154] In order to further verify the effectiveness of the technology proposed in this invention in the complex speed conditions of permanent magnet motors, two variable speed conditions are used for experiments. The variable speed conditions are divided into acceleration and deceleration, and the corresponding speed waveforms are as follows: Fig.13 shown.

[0155] Depend on Fig.13 (a) and Fig.13 (b) It can be seen that the speed signal is decomposed using the empirical Fourier decomposition algorithm adopted by the present invention to extract speed fault harmonics under different variable speed conditions, and then the fault harmonics are resampled using the rotor position angle calculated by the stator current fundamental wave to perform fault diagnosis, and the following is obtained: Fig.14 and Fig.15 Diagnostic results of outer race fault and inner race fault under two working conditions.

[0156] Depend on Fig.14 (a) and Fig.14 (b) and Fig.15 (a) and Fig.15 (b) It can be seen that characteristic orders related to outer ring faults and inner ring faults can be found in the order diagram, and their values ​​are close to the theoretical values. Therefore, it can be proved that the technology proposed in the present invention is capable of diagnosing faults under variable speed conditions. At the same time, the present invention can solve the dependence on working conditions and is more suitable for actual working conditions.

[0157] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0158] In addition, it should be noted that the scope of the methods and systems in the embodiments of the present invention is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.

[0159] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A permanent magnet motor fault diagnosis method based on multi-electrical signal fusion, characterized in that: The steps include: Step S1, calculating the fault characteristic frequency in the speed signal when the bearing fails; Step S2, collecting the stator current signal and the speed signal in the permanent magnet motor control system; Step S3, using an empirical Fourier decomposition algorithm to decompose the collected stator current signal and speed signal, and extracting fault harmonics and current fundamental waves from the decomposed components; Step S4, determining whether the motor is running at a constant speed, if so, executing step S5; If not, execute step S6; Step S5, performing envelope spectrum analysis on the fault harmonics to obtain a frequency spectrum, and obtaining frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination; Step S6, using the extracted current fundamental wave to obtain rotor phase information to obtain a continuous angle increment signal of the rotor; Step S7, using the obtained continuous angle increment signal to perform angle domain resampling on the extracted fault harmonics, so as to convert the time domain signal into an angle domain signal; Step S8, performing envelope spectrum analysis on the resampled angular domain signal, and determining whether a fault occurs and the type of the fault by comparing the fault information in the spectrum with the theoretical value.

2. The method according to claim 1, characterized in that In step S1, the fault characteristic frequency includes the outer ring fault characteristic frequency and the inner ring fault characteristic frequency, which is expressed by the following formula: Among them, f ORF 、f IRF They are the characteristic frequency of outer ring fault and inner ring fault respectively; FCO ORF 、FCO IRF分 They are the characteristic frequency coefficient of outer ring fault and the characteristic frequency coefficient of inner ring fault respectively; N B is the number of bearing balls; D b D is the ball diameter; p is the ball diameter; α is the bearing contact angle; f r is the mechanical rotation frequency of the motor.

3. The method according to claim 1, characterized in that In step S2, the stator current signal and the speed signal are obtained through the current sensor and the incremental encoder in the permanent magnet motor control system respectively.

4. The method according to claim 3, characterized in that In step S6, the extracted current fundamental wave i(t) is used to obtain the rotor phase information to obtain the continuous angle increment signal C of the rotor. φ (t), including: The current fundamental wave i(t) consists of two parts: amplitude component and phase component, which can be expressed as follows: Among them, f i (t) is the amplitude part; is the phase part, where j is the phase unit, It is obtained by Hilbert transform, as shown below: H represents the Hilbert transform, τ is the integral variable, and the instantaneous phase φ(t) of the current fundamental wave i(t) is calculated by the following formula: The continuous angle increment signal C of the rotor is obtained by processing φ(t) through the accumulation function unwrap() in Matlab. φ (t).

5. The method according to claim 4, characterized in that In step S7, the angular domain resampling is expressed by the following formula: Rf(θ)=resample(f(t),C φ (t)) (17) Wherein, resample() is an existing function in MATLAB, and f(t) is the fault harmonic.

6. A permanent magnet motor fault diagnosis system based on multi-electrical signal fusion for implementing the method according to any one of claims 1 to 5, characterized in that: include: A fault characteristic frequency calculation module, which is used to calculate the fault characteristic frequency in the speed signal when the bearing fails; A stator current signal and speed signal acquisition module, which is used to acquire the stator current signal and speed signal in the permanent magnet motor control system; A fault harmonic and current fundamental wave extraction module is used to decompose the collected stator current signal and speed signal using an empirical Fourier decomposition algorithm, and extract the fault harmonic and current fundamental wave from the decomposed components; A working condition judgment module is used to judge whether the motor is running under a constant speed condition; A fault diagnosis and type discrimination module is used to perform envelope spectrum analysis on fault harmonics to obtain a frequency spectrum, and obtain frequency characteristic components related to the fault to achieve fault diagnosis and type discrimination; An angle increment signal acquisition module, which is used to obtain rotor phase information using the extracted current fundamental wave to obtain a continuous angle increment signal of the rotor; An angular domain signal conversion module, which is used to perform angular domain resampling on the extracted fault harmonics using the obtained continuous angular increment signal, so as to convert the time domain signal into an angular domain signal; The fault determination and type discrimination module is used to perform envelope spectrum analysis on the resampled angular domain signal, and to determine whether a fault has occurred and to discriminate the fault type by comparing the fault information in the spectrum with the theoretical value.

Images