A method for detecting broken rotor bar fault of asynchronous motor without voltage flicker

Abstract

A method for detecting a broken rotor bar fault of an asynchronous motor that can avoid the influence of voltage flicker is disclosed. The detection method is for a three-phase cage-type asynchronous motor. The detection device used includes a voltage regulator, a current transformer, a voltage transformer, a data collector, and a computer. During the detection, when the power supply voltage flicker exists and affects the detection of the broken rotor bar fault, the stator current and voltage of the asynchronous motor are measured and analyzed, the main characteristic components of the broken rotor bar fault are investigated, the main characteristic frequency, auxiliary characteristic frequency, and power supply voltage flicker characteristic frequency of the broken rotor bar fault are calculated and compared, the characteristic frequency and component of the broken rotor bar fault are selected, and then the characteristic components are analyzed to accurately determine whether the broken rotor bar fault occurs. The present invention can accurately detect the broken rotor bar fault of an asynchronous motor, is immune to the adverse effects of the power supply voltage flicker, thereby avoiding misjudgment, and has good practical value and promotion prospects.

Description

Technical Field

[0001] The invention relates to the technical field of measuring electric and magnetic variables, and in particular to a detection technology for a rotor broken bar fault of a three-phase cage asynchronous motor. Background Art

[0002] During the operation of the three-phase cage-type asynchronous motor, the rotor bars are subjected to alternating stresses such as radial electromagnetic force, rotational electromagnetic force, centrifugal force, thermal bending deflection force, etc., and rotor manufacturing defects may lead to broken bar failure, the probability of which is about 15%.

[0003] Broken rotor bars are a typical progressive fault. Usually one or two bars are broken at the beginning, and then gradually develop to the point where the output of the asynchronous motor decreases or even stops. Therefore, it is necessary to implement broken rotor bar fault detection.

[0004] After a three-phase cage asynchronous motor has a broken rotor bar fault, a specific component of the frequency determined by equation (1) will appear in its stator current, which is usually called the broken rotor bar fault characteristic component. In the present invention, it is called the main characteristic component of the broken rotor bar fault. The corresponding frequency is is called the main characteristic frequency of the rotor bar broken fault. In formula (1), s Indicates the slip rate; f 1 represents the power supply frequency, that is, the fundamental frequency.

[0005] Therefore, by performing spectrum analysis on the stator current signal and extracting and examining the characteristic component, the rotor bar fault detection can be realized. This is the rotor bar fault detection method based on stator current signal analysis, usually called MCSA (Motor Current Signal Analysis) method, that is, asynchronous motor current signal analysis method.

[0006] The original rotor broken bar fault detection method is to directly perform FFT (Fast Fourier Transform) spectrum analysis on the steady-state stator current signal, and then determine whether (1 ± 2 s ) f 1 frequency component and its amplitude to determine whether the rotor is broken. When the rotor is slightly broken, (1 ± 2 s ) f The magnitude of component 1 is relative to f 1 component is very small, and the slip rate of the asynchronous motor is s Very small, (1 ± 2s ) f 1 and f 1These two frequency values ​​are close. If we do FFT spectrum analysis directly, then (1 ± 2 s ) f 1 portion may be f 1 component of the leak was overwhelmed.

[0007] To solve this problem, some people have proposed to introduce adaptive filtering to highlight the characteristic components of rotor bar failure in the spectrum diagram. ± 2 s ) f 1 component; and then use refined FFT spectrum analysis to improve the recognition of spectrum analysis. Therefore, the detection performance of rotor broken bar fault is greatly improved.

[0008] In addition, the broken rotor bar fault will also cause the stator current to appear in the formula (1 ±4 s ) f 1 and The specific component of the determined frequency is referred to as the auxiliary characteristic component of the rotor bar fault in the present invention, and the corresponding frequency is referred to as the auxiliary characteristic frequency of the rotor bar fault. However, since the auxiliary characteristic component of the rotor bar fault is significantly smaller in amplitude than the main characteristic component of the rotor bar fault, it has not received widespread attention in the practice of rotor bar fault detection.

[0009] In summary, although the detection method of broken rotor bar fault of three-phase cage asynchronous motor has made great progress, there are still "misjudgments" in the practice of broken rotor bar fault detection. 2 s f 1) Fluctuating loads will generate specific components of certain frequencies in the stator current, an effect similar to a broken rotor bar fault, causing the MCSA method to mistakenly identify load fluctuations as a broken rotor bar fault. For another example, the rotor axial wind duct will cause magnetic field asymmetry, which is similar to the magnetic effect of a broken rotor bar fault, and therefore may also cause the MCSA method to "misjudge". For another example, when an asynchronous motor drives mechanical equipment such as pumps, fans, and compressors, if the number of magnetic poles of the asynchronous motor is equal to an integer multiple of the number of blades of the mechanical equipment, the blade through-frequency vibration will generate specific components of certain frequencies in the stator current, an effect similar to a broken rotor bar fault, and thus may also cause the MCSA method to "misjudge".

[0010] In addition, the three-phase cage-type asynchronous motor runs under the excitation of the three-phase supply voltage. Similar to the above factors, the quality of the supply voltage can also have an adverse effect on the detection of broken rotor bar faults.

[0011] In 1994, the International Electrotechnical Commission listed voltage flicker as one of the important indicators for measuring power quality. Since then, the problem of voltage flicker has received continuous attention. In recent years, renewable energy such as wind power and photovoltaic power have been connected to the power grid on a large scale, and their inherent volatility has also exacerbated the problem of voltage flicker.

[0012] In fact, similar to factors such as load fluctuations, axial air ducts, and blade through-frequency vibration, power supply voltage flicker can also lead to "misjudgment" of the MCSA method, which is an important issue that needs to be solved urgently.

[0013] The Chinese patent application number CN201710761248.3 and the name "A method for judging the broken bar fault of an asynchronous motor rotor" discloses a detection method, which uses the following steps to detect broken bars: the collector collects voltage and current data through the current clamp and the voltage clamp, and performs current spectrum analysis. If the current signal is normal, the asynchronous motor rotor has no broken bar fault. If the current signal is abnormal, a sideband will appear at the position that differs from the power supply frequency by two times the slip frequency, and then the next step of processing is performed to analyze the voltage spectrum. If the voltage signal is normal, it is determined that the asynchronous motor rotor has no broken bars. If the voltage signal is abnormal, a sideband appears at the position that differs from the power supply frequency by two times the slip frequency, indicating that the asynchronous motor rotor has a broken bar fault. This patent is aimed at the existing technology that only uses the asynchronous motor current signal analysis when judging the broken rotor bar fault, which will result in misjudgment. Therefore, after improvement, the stator current and stator voltage are judged in sequence to detect the broken rotor bar fault, thereby reducing the misjudgment rate. Although this patent reduces the misjudgment rate of rotor bar breakage detection, in actual field operation, the power supply voltage flicker still causes signal interference, resulting in misjudgment of rotor bar breakage, especially when the power supply voltage flicker and rotor bar breakage exist at the same time. In other words, this patent does not solve the problem of the impact of power supply voltage flicker on rotor bar breakage fault detection.

[0014] In summary, when detecting a broken rotor bar fault in a three-phase cage-type asynchronous motor, in addition to the abnormal measurement results caused by the broken rotor bar, the abnormal mechanical vibration of the asynchronous motor and the flicker of the power supply voltage will have an adverse effect on the detection of the broken rotor bar fault and affect the accuracy of the detection results. Therefore, the industry is in urgent need of proposing a method for detecting a broken rotor bar fault in a cage-type asynchronous motor that can prevent interference from power supply voltage flicker. Summary of the invention

[0015] The purpose of the present invention is to provide an asynchronous motor rotor broken bar fault detection method that can avoid the influence of voltage flicker. The method is immune to the adverse influence of power supply voltage flicker and can accurately determine the rotor broken bar fault. The detection device of the present invention has a simple structure and is easy to operate. When calculating and analyzing the stator current signal, the analysis method adopted is reliable and accurate. The main characteristic frequency components, auxiliary characteristic frequency components and amplitude and frequency of the power supply voltage flicker component of the asynchronous motor rotor broken bar fault can be quickly detected, thereby avoiding misjudgment.

[0016] The objective of the present invention is achieved through the following technical solutions:

[0017] A method for detecting a broken bar fault of an asynchronous motor rotor that can avoid the influence of voltage flicker. The asynchronous motor targeted by the detection method is a three-phase cage-type asynchronous motor. The detection device used includes a voltage regulator, a current transformer, a voltage transformer, a data collector, a computer, a load, and a tachometer. The voltage regulator is used to adjust the test voltage of the asynchronous motor D under test, the current transformer and the voltage transformer are used to collect the feeding current and voltage, the data collector sends the measured data to the computer for calculation and analysis, the tachometer is used to measure the rotation speed of the asynchronous motor, and the load is connected to the output shaft of the asynchronous motor.

[0018] During detection, when the power supply voltage flicker exists and affects the detection of the broken rotor bar fault, the stator current and voltage of the asynchronous motor are measured and analyzed, the main characteristic components of the broken rotor bar fault are examined, the main characteristic frequency, auxiliary characteristic frequency and supply voltage flicker characteristic frequency of the broken rotor bar fault are calculated and compared, the characteristic frequency and components unique to the broken rotor bar fault are selected, and then the characteristic components are analyzed to accurately determine whether the broken rotor bar fault occurs.

[0019] The above asynchronous motor rotor broken bar fault detection method that can avoid the influence of voltage flicker is specifically detected and analyzed in the following steps:

[0020] a. Use a rotor bar fault detection method based on stator current signal analysis to measure the stator current signal of one phase and use an adaptive filtering method to offset the stator current f 1 frequency component, and perform refined FFT spectrum analysis to obtain refined FFT spectrum. ± 2 s ) f 1. Whether the main characteristic component of the rotor broken bar fault of the determined frequency exists, predict whether the rotor state has a fault: if the spectrum peak corresponding to the main characteristic component of the rotor broken bar fault appears in the refined FFT spectrum, the rotor is broken; if the spectrum peak does not exist in the refined FFT spectrum, the rotor is healthy;

[0021] b. Detect power supply voltage flicker and determine whether it affects rotor bar fault detection

[0022] First, according to step a, f 1. Calculate the single-phase supply voltage u s ( t )exist f 1 to determine the amplitude of its fundamental component; then, u s ( t ) to filter out its fundamental component by adaptive filtering, and perform detailed FFT spectrum analysis to determine the amplitude and frequency of the power supply voltage flicker component f m ; Then, according to f m Is it close to 2? s f 1. Determine whether the power supply voltage flicker affects the rotor bar fault detection. If f m Not close to 2 s f 1, the power supply voltage flicker does not affect the rotor bar fault detection, the prediction of the rotor state in step a is established, and the rotor bar fault detection is completed; otherwise, the power supply voltage flicker affects the rotor bar fault detection, and the subsequent steps need to be further performed;

[0023] c. Determine the characteristic frequency of the current rotor bar failure, which is unique to the current rotor bar failure and different from the supply voltage flicker, and record it as f RF .

[0024] First, calculate the main characteristic frequency and auxiliary characteristic frequency of the rotor broken bar fault. The calculation formula is as follows:

[0025] ;

[0026] Calculate the characteristic frequency of the power supply voltage flicker close to the main and auxiliary characteristic frequencies of the broken rotor bar fault. The calculation formula is as follows:

[0027] ;

[0028] Calculate the characteristic frequency of power supply voltage flicker close to the characteristic frequency of rotor bar fault. "Close" means that the characteristic frequency of power supply voltage flicker is within the range of [-5Hz — +5Hz] of the characteristic frequency of rotor bar fault.

[0029] Then, the characteristic frequency of the broken rotor bar fault is compared with the characteristic frequency of the power supply voltage flicker, so as to select the characteristic frequency unique to the broken rotor bar fault.f RF ;

[0030] d. Check the refined FFT spectrum of the one-phase stator current signal obtained in step a. f RF Whether there is a corresponding spectrum peak nearby, and make further accurate judgment: If the spectrum graph is f RF If there is no spectrum peak near f RF If a spectrum peak appears near the rotor bar, it means that the rotor bar is broken, accompanied by power supply voltage flicker. Beneficial Effects

[0031] The most significant advantage of the present invention is that it can accurately detect the broken rotor bar fault of the asynchronous motor, effectively avoiding the adverse effects caused by the flicker of the power supply voltage, thereby avoiding misjudgment. In recent years, renewable energy sources such as wind power and photovoltaic power have been connected to the power grid on a large scale. Their inherent volatility has brought about serious power supply voltage flicker problems, resulting in a large number of misjudgments in the existing broken rotor bar fault detection methods. Therefore, the present invention overcomes these defects of the existing technology and has good practical value and broad promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] Figure 1 is a schematic diagram of the structure of the detection device;

[0034] Figure 2 It is a flowchart of the present invention;

[0035] Figure 3 It is the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the supply voltage is normal and the rotor is healthy;

[0036] Figure 4 It is the transient waveform of one-phase power supply voltage when the power supply voltage is normal;

[0037] Figure 5 It is the refined FFT spectrum within the (40~60) Hz range of one-phase power supply voltage when the power supply voltage is normal;

[0038] Figure 6 It is the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the power supply voltage is normal and the rotor bar is broken;

[0039] Figure 7 It is the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the supply voltage flickers and the rotor is healthy;

[0040] Figure 8 It is the transient waveform of one-phase power supply voltage when the power supply voltage flickers;

[0041] Fig. 9 It is the refined FFT spectrum within the (40~60) Hz range of one-phase power supply voltage when the power supply voltage flickers;

[0042] Fig.10 It is the refined FFT spectrum of one-phase stator current in the range of (335~345) Hz when the supply voltage flickers and the rotor is healthy;

[0043] Fig.11 It is the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the power supply voltage flickers and the rotor bar is broken;

[0044] Fig.12 It is the refined FFT spectrum of the single-phase stator current in the range of (335~345) Hz when the power supply voltage flickers and the rotor bar is broken.

[0045] The meanings of the symbols used in this paper are: FFT, Fast Fourier Transform; f RF , the characteristic frequency of rotor bar failure, which is unique to it and different from the supply voltage flicker, s , slip rate; f 1. Power supply frequency (fundamental frequency); MCSA, Motor Current Signal Analysis; u s ( t ), one-phase supply voltage; A 1. u s ( t )’s fundamental component amplitude; A m , the amplitude of the amplitude modulated wave voltage; f m , the frequency of the amplitude modulated wave voltage, that is, the power supply voltage flicker frequency; ϕ m , the initial phase angle of the amplitude modulated wave voltage; t ,time; m = k ( Z 2 / P )±1, where k = 0, 1, 2, …; Z 2. Number of rotor slots; P , pole pair number; ν = 6 c +1, among which c=0, ±1, ±2, ….

[0046] The component numbers in the figure are: 1. voltage regulator, 2. current transformer, 3. voltage transformer, 4. data acquisition device, 5. computer, 6. load, 7. tachometer, D. asynchronous motor. DETAILED DESCRIPTION

[0047] The present invention proposes a relatively complete detection method for a broken bar fault of an asynchronous motor rotor, which is immune to the adverse effects of power supply voltage flicker and can accurately determine a broken bar fault of the rotor. The detection device of the present invention has a simple structure and is easy to operate. When calculating and analyzing the stator current signal, the analysis method adopted is reliable and accurate. The main characteristic frequency components, auxiliary characteristic frequency components and amplitude and frequency of the power supply voltage flicker component of the broken bar fault of the asynchronous motor rotor can be quickly detected, and comparison and analysis can be made to avoid misjudgment.

[0048] In the present invention, the power supply voltage flicker is regarded as the modulation of the fundamental wave component of the power supply voltage by an amplitude modulated wave voltage. Taking a single-phase power supply voltage as an example, its mathematical model is defined as formula (4).

[0049] ;

[0050] In formula (4), u s ( t ) represents a single-phase supply voltage; A 1 means u s ( t )’s fundamental component amplitude; A m , f m , ϕ m Respectively represent the amplitude, frequency and initial phase angle of the amplitude modulated wave voltage; t Indicates time.

[0051] According to the national standard GB / T12326-2008 "Power Quality Voltage Fluctuation and Flicker", the amplitude of the amplitude modulated wave voltage is much smaller than the fundamental component of the power supply voltage. For example, for a power supply system with a nominal voltage level below 1kV, its voltage fluctuation limit is Not more than 4%, therefore A m / A 1<1.4%. In addition, the frequency of the amplitude modulated wave voltage f m Between 0.05~35 Hz.

[0052] Assuming that the rotor quality is intact, and the supply voltage has a flicker relative to the fundamental component as shown in equation (4), ignoring nonlinear factors such as asynchronous motor saturation, the supply voltage can be approximately regarded as the fundamental component and ( f 1± f m ) components. In the present invention, the power supply voltage ( f 1± f m ) component is called the power supply voltage flicker component, and f m It is called the supply voltage flicker frequency.

[0053] According to the theory of rotating electromagnetic field of asynchronous motor, the supply voltage ( f 1+ f m ) The frequency expression of the stator current component caused by the flicker component is shown in formula (5); similarly, the supply voltage ( f 1- f m The frequency expression of the stator current component caused by the flicker component is shown in formula (6), that is, the supply voltage flicker component will cause a specific component of the frequency determined by formulas (5) and (6) to appear in the stator current, which is called the supply voltage flicker characteristic component in the present invention, and the corresponding frequency is called the supply voltage flicker characteristic frequency.

[0054] ;

[0055] In formula (5) and (6), m = k ( Z 2 / P )±1( k = 0, 1, 2, …; Z 2 indicates the number of rotor slots; P represents the number of pole pairs); ν = 6 c +1 c = 0, ±1, ±2, …).

[0056] According to equations (5) and (6), the value of the characteristic frequency of power supply voltage flicker is very complex, involving m , ν, s as well as f m For the fundamental wave only ( m = 1, ν = 1), the flicker characteristic frequency is f 1± f m , that is, the stator current containsf 1± f m The flicker characteristic component of the frequency.

[0057] According to formula (1), the frequency of the main characteristic component of the rotor bar broken fault is (1 ± 2 s ) f 1 = f 1±2 sf 1. Usually when the asynchronous motor is running, s Between 0.002 and 0.05, then 2 sf 1Between 0.2 and 5 Hz (select f 1 = 50 Hz). As mentioned before, f m Between 0.05~35 Hz (see national standard GB / T12326-2008). f m It may be equal to 2 sf 1, at this time, the power supply voltage flicker characteristic component is—— f 1± f m Frequency components, which may confuse the main characteristic components of rotor broken bar faults—— f 1±2 sf 1 frequency component, thus affecting the detection of rotor bar failure, causing the MCSA method to "misjudge" the power supply voltage flicker to be mistakenly judged as a rotor bar failure.

[0058] By analyzing why power supply voltage flicker affects rotor broken bar fault detection, we have found a way to prevent misjudgment.

[0059] Figure 1 What is drawn is the detection device of the present invention, and the method of the present invention can also be verified by using the device. The verification adopts an experimental asynchronous motor to detect the broken rotor bar fault. The asynchronous motor in the figure adopts a Y100L-2 three-phase cage-type asynchronous motor (3kW, 380V, 50Hz). In addition to the normal rotor, a faulty rotor is equipped to simulate the broken bar fault. The faulty rotor has a broken conductor bar (drilled 10mm away from the end ring, with a diameter of 6mm and a depth of 10mm). The voltage regulating device is a distribution transformer (which applies the actual grid voltage to the asynchronous motor) and a programmable AC power supply (which simulates the flickering power supply voltage). In the experiment, the asynchronous motor load is approximately full load and no adjustment is made in order to keep it constant, s Approx. 0.027 6. In the experiment, two measurements are required. The first one is to test the asynchronous motor after the normal rotor is installed, and the second one is to test the asynchronous motor after the faulty rotor is installed.

[0060] 1. Normal power supply voltage

[0061] In order to combine with reality, the actual grid voltage is applied to the asynchronous motor through the distribution transformer. The actual grid voltage has no flicker but contains harmonics.

[0062] For a healthy rotor, experiments are conducted and one-phase stator current signal and one-phase supply voltage signal are measured.

[0063] Implement step a of the present invention (see Figure 2 ), the MCSA method based on adaptive filtering and refined FFT spectrum analysis is used to obtain the refined FFT spectrum of the one-phase stator current signal, as shown in Figure 3 shown. Figure 3 That is, the refined FFT spectrum of the stator current of one phase in the range of (40~60) Hz when the power supply voltage is normal and the rotor is healthy. f 1 = 50.04 Hz, s ≈ 0.0276, while the main characteristic frequency of rotor bar failure (1-2 s ) f 1≈ 47.28 Hz, (1+2 s ) f 1≈ 52.80 Hz. Figure 3 It can be seen that there is no obvious spectral peak near the main characteristic frequency of the broken rotor bar fault, so it can be predicted that the rotor is healthy.

[0064] Implement step b of the present invention (see Figure 2 ), for one-phase power supply voltage signal u s ( t )( Figure 4 Represents its transient waveform), and uses a method based on adaptive filtering and refined FFT spectrum analysis to obtain its refined FFT spectrum, as shown in Figure 5 shown. Figure 5 That is, the refined FFT spectrum of one-phase power supply voltage within the range of (40~60) Hz when the power supply voltage is normal. Figure 5 It can be determined that the power supply voltage flicker does not exist. Therefore, it can be confirmed that the rotor is healthy and the rotor broken bar fault detection process is completed. Obviously, the test result is accurate.

[0065] The normal rotor is replaced by a broken-bar rotor, and the experiment is carried out to measure the one-phase stator current signal and the one-phase supply voltage signal.

[0066] Implement step a of the present invention (see Figure 2 ), the MCSA method based on adaptive filtering and refined FFT spectrum analysis is used to obtain the refined FFT spectrum of the one-phase stator current signal, such as Figure 6 shown. Figure 6 That is, the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the power supply voltage is normal and the rotor bar is broken. f 1 = 50.04 Hz, s ≈ 0.0276, while the main characteristic frequency of rotor bar failure (1-2 s ) f 1≈ 47.28 Hz, (1+2 s ) f 1≈ 52.80 Hz. Figure 6 It can be seen that there is an obvious spectrum peak near the main characteristic frequency of the rotor bar broken fault, so it can be predicted that the rotor bar is broken.

[0067] Implement step b of the present invention (see Figure 2 ), for one-phase power supply voltage signal u s ( t ), using a method based on adaptive filtering and refined FFT spectrum analysis to obtain its refined FFT spectrum, such as Figure 5 As shown. Figure 5 It can be determined that the power supply voltage flicker does not exist. Therefore, it can be confirmed that the rotor bar is broken and the rotor bar fault detection process is completed. Obviously, the detection result is accurate.

[0068] 2. Power supply voltage flicker

[0069] The three-phase voltage signal applied to the asynchronous motor by the distribution transformer is measured and stored, and the flicker component is superimposed according to formula (4). The supply voltage containing the flicker component is generated by the programmable AC power supply and applied to the asynchronous motor. The parameters of the supply voltage flicker are set as: A m ≈ 0.014 A 1 ≈ 4.16 V, f m ≈ 2 s f 1 = 2.76 Hz. NOTE: Setting f m with 2 s f 1 is approximately equal, which is the most unfavorable situation in which the power supply voltage flicker affects the detection of rotor broken bar fault.

[0070] For a healthy rotor, experiments are conducted and one-phase stator current signal and one-phase supply voltage signal are measured.

[0071] Implement step a of the present invention (see Figure 2 ), the MCSA method based on adaptive filtering and refined FFT spectrum analysis is used to obtain the refined FFT spectrum of the one-phase stator current signal, as shown in Figure 7 shown. Figure 7 This is the refined FFT spectrum of the stator current of one phase in the range of (40~60) Hz when the power supply voltage flickers and the rotor is healthy. f 1 = 50.04 Hz, s ≈ 0.0276, while the main characteristic frequency of rotor bar failure (1-2 s ) f 1≈ 47.28 Hz, (1+2 s ) f 1≈ 52.80 Hz. Figure 7 It can be seen that there is an obvious spectrum peak near the main characteristic frequency of the rotor bar fault, so it can be predicted that the rotor bar is broken. Of course, the actual situation is that the rotor is healthy, so the prediction result is wrong, which is caused by the power supply voltage flicker. This shows that the existing MCSA method is affected by the power supply voltage flicker, and thus the power supply voltage flicker is "misjudged" as a rotor bar fault.

[0072] Implement step b of the present invention (see Figure 2 ), for one-phase power supply voltage signal u s ( t )( Figure 8 Represents its transient waveform), and uses a method based on adaptive filtering and refined FFT spectrum analysis to obtain its refined FFT spectrum, as shown in Fig. 9 shown. Fig. 9 That is, the refined FFT spectrum of one-phase power supply voltage in the range of (40~60) Hz when the power supply voltage flickers. Fig. 9 It can be seen that the supply voltage signal u s ( t )'s refined FFT spectrum is in ( f 1- f m ) = 47.35 Hz and ( f 1+ fm ) = 52.80 Hz, so it can be determined that the power supply voltage flicker exists; and, f m ≈ 2 s f 1 = 2.76 Hz, which affects the detection of rotor broken bar faults. For ease of understanding, Fig. 9 and Figure 5 In addition, Figure 8 and Figure 4 By comparison, it can be seen that the power supply voltage flicker u s ( t ) The envelope of the transient waveform pulsates. Therefore, it is necessary to perform the subsequent steps c and d.

[0073] Implement step c of the present invention (see Figure 2 ), according to formula (1) and formula (2) and (3), the main and auxiliary characteristic frequencies of the rotor bar fault are calculated, as shown in Table 1. According to formula (5) and (6), the characteristic frequencies of power supply voltage flicker close to the main and auxiliary characteristic frequencies of the rotor bar fault are calculated, 47.28 Hz, 52.80 Hz, 339.23 Hz, and 344.75 Hz. "Close" means that the characteristic frequency of power supply voltage flicker is in the range of [characteristic frequency of rotor bar fault - 5Hz characteristic frequency of rotor bar fault + 5Hz], for example, [42.28Hz 52.28Hz], [47.80Hz 57.80Hz]. By comparing the above calculation results, the characteristic frequency unique to the current rotor bar fault and different from the power supply voltage flicker can be selected. f RF , which are 44.52 Hz, 55.56 Hz, and 341.99 Hz.

[0074] In other words, f RF is the characteristic frequency of the rotor bar fault, which is some of the values ​​in the multiple calculation results of the above three formulas. Table 1 below is the characteristic frequency of the rotor bar fault calculated according to the above three formulas, namely 47.28Hz, 52.80Hz, 44.52Hz, 55.56Hz, 339.23Hz, and 341.99Hz; Formula 2 is the characteristic frequency of the power supply voltage flicker, which also has multiple calculation results, but f RF Not included in it, such as the characteristic frequencies of power supply voltage flicker corresponding to Table 1 are 47.28 Hz, 52.80 Hz, 339.23 Hz, and 344.75 Hz. Therefore, f RFIt is the "unique" characteristic frequency of the rotor bar failure, which is unrelated to the power supply voltage flicker. By comparing the above calculation results, we can select f RF They are 44.52 Hz, 55.56 Hz, and 341.99 Hz.

[0075] Step d of implementing the present invention (see Figure 2 ), check the refined FFT spectrum of the one-phase stator current signal obtained in step a. f RF Check whether there is a corresponding spectrum peak nearby, and judge the rotor status again - healthy, and the power supply voltage flickers; or faulty, and the power supply voltage flickers. Figure 7 It can be seen that the refined FFT spectrum of one-phase stator current is f RF There are no peaks near (44.52 Hz, 55.56 Hz). In addition, Fig.10 This is the refined FFT spectrum of the stator current of one phase in the range of (335~345) Hz when the power supply voltage flickers and the rotor is healthy. f RF There is no spectrum peak near (341.99 Hz). Therefore, it can be judged that the rotor is healthy, but the power supply voltage flickers, and the rotor broken bar fault detection process is completed. Obviously, the detection result of the present invention is accurate.

[0076] The normal rotor is replaced by a broken-bar rotor, and the experiment is carried out to measure the one-phase stator current signal and the one-phase supply voltage signal.

[0077] Implement step a of the present invention (see Figure 2 ), the MCSA method based on adaptive filtering and refined FFT spectrum analysis is used to obtain the refined FFT spectrum of the one-phase stator current signal, as shown in Fig.11 shown. Fig.11 That is, the refined FFT spectrum of one-phase stator current in the range of (40~60) Hz when the power supply voltage flickers and the rotor bar is broken. f 1 = 50.04 Hz, s ≈0.027 6, while the main characteristic frequency of rotor bar failure (1-2 s ) f 1≈ 47.28 Hz, (1+2 s ) f 1≈ 52.80 Hz. Fig.11 It can be seen that there is an obvious spectrum peak near the main characteristic frequency of the rotor bar broken fault, so it can be predicted that the rotor bar is broken.

[0078] Implement step b of the present invention (see Figure 2 ), for one-phase power supply voltage signal u s ( t )( Figure 8 Represents its transient waveform), and uses a method based on adaptive filtering and refined FFT spectrum analysis to obtain its refined FFT spectrum, as shown in Fig. 9 shown. Fig. 9 That is, the refined FFT spectrum of one-phase power supply voltage in the range of (40~60) Hz when the power supply voltage flickers. Fig. 9 It can be seen that the supply voltage signal u s ( t )'s refined FFT spectrum is in ( f 1- f m ) = 47.35 Hz and ( f 1+ f m ) = 52.80 Hz, so it can be determined that the power supply voltage flicker exists; and, f m ≈ 2 s f 1 = 2.76 Hz, which affects the detection of rotor broken bar fault. Therefore, it is necessary to perform the subsequent steps c and d.

[0079] Implement step c of the present invention (see Figure 2 ), according to formula (1) and formula (2) and (3), the main and auxiliary characteristic frequencies of the rotor bar fault are calculated, as shown in Table 1. According to formula (5) and (6), the characteristic frequencies of power supply voltage flicker close to the main and auxiliary characteristic frequencies of the rotor bar fault are calculated, 47.28 Hz, 52.80 Hz, 339.23 Hz, and 344.75 Hz. By comparing the above calculation results, the characteristic frequency unique to the current rotor bar fault and different from the power supply voltage flicker can be selected. f RF , which are 44.52 Hz, 55.56 Hz, and 341.99 Hz.

[0080] Step d of implementing the present invention (see Figure 2 ), check the refined FFT spectrum of the one-phase stator current signal obtained in step a. f RF Check whether there is a corresponding spectrum peak nearby, and judge the rotor status again - healthy, and the power supply voltage flickers; or faulty, and the power supply voltage flickers. Fig.11 It can be seen that the refined FFT spectrum of one-phase stator current is f RFThere are spectrum peaks near (44.52 Hz, 55.56 Hz). In addition, Fig.12 This is the refined FFT spectrum of the stator current of one phase in the range of (335~345) Hz when the power supply voltage flickers and the rotor bar is broken. f RF A spectrum peak appears near (341.99 Hz). Therefore, it can be determined that the rotor bar is broken and the power supply voltage flickers, and the rotor bar fault detection process is completed. Obviously, the detection result of the present invention is accurate.

[0081] The above experiments show that the power supply voltage flicker does have an adverse effect on the detection of rotor bar failure and causes the existing MCSA method to misjudge. The method provided by the present invention can avoid the adverse effect of power supply voltage flicker, which is a significant advantage of the present invention. Compared with the existing MCSA method, the present invention has made significant improvements, which are explained as follows.

[0082] According to the theory of rotating electromagnetic field of asynchronous motor, the present invention derives the frequency expression of the stator current component caused by the power supply voltage flicker component, as shown in equations (5) and (6). In the case where the power supply voltage flicker exists and affects the detection of the rotor bar fault, the characteristic frequency of the power supply voltage flicker can be calculated according to the expression, and then compared with the main and auxiliary characteristic frequencies of the rotor bar fault to select the characteristic frequencies and components that are unique to the rotor bar fault and different from the power supply voltage flicker. In the case where the power supply voltage flicker exists and affects the detection of the rotor bar fault, the present invention not only examines the main characteristic components of the rotor bar fault, but also examines the characteristic components that are unique to the rotor bar fault and different from the power supply voltage flicker, so as to provide accurate detection results to avoid the adverse effects of the power supply voltage flicker.

[0083] In summary, when performing detection and analysis, the present invention first performs step a to predict the rotor state - healthy or faulty, and then determines the rotor state according to step a. f 1. Calculate the single-phase supply voltage u s ( t )exist f 1 to determine the amplitude of its fundamental component; then, u s ( t ) to filter out its fundamental component by adaptive filtering, and perform detailed FFT spectrum analysis to determine the amplitude and frequency of the power supply voltage flicker component f m ; Then, according to f m Is it close to 2? s f1. Determine whether the power supply voltage flicker affects the rotor bar fault detection. If the power supply voltage flicker does not affect the rotor bar fault detection, the prejudgment of the rotor state in step a is established, and the rotor bar fault detection is completed; otherwise, the subsequent steps need to be further performed.

[0084] Determine the characteristic frequency of the current rotor bar failure, which is unique to the current rotor bar failure and different from the power supply voltage flicker, and record it as f RF .

[0085] Specifically, the main and auxiliary characteristic frequencies of the rotor bar fault are calculated according to equations (1), (2) and (3), and the characteristic frequencies of the power supply voltage flicker close to them are calculated according to equations (5) and (6). The two are compared to select the characteristic frequency unique to the rotor bar fault. f RF It should be noted that the characteristic frequency of the power supply voltage flicker close to the characteristic frequency of the rotor bar fault is calculated. "Close" means that the characteristic frequency of the power supply voltage flicker is within the range of [characteristic frequency of the rotor bar fault - 5Hz - characteristic frequency of the rotor bar fault + 5Hz]. This is because the main and auxiliary characteristic frequencies of the rotor bar fault are close to 2 sf 1 is closely related to the asynchronous motor when it is running 2 sf 1Between 0.2 and 5 Hz (select f 1 = 50 Hz), the upper limit is 5 Hz for reliability.

[0086] Then, check the refined FFT spectrum of the one-phase stator current signal obtained in step a to see if f RF Check whether there is a corresponding spectrum peak nearby and judge the rotor status again - healthy, with power supply voltage flickering; or faulty, with power supply voltage flickering.

[0087] If the spectrum is f RF If there is no spectrum peak near f RF If a spectrum peak appears near the rotor bar, the rotor bar is broken and there is supply voltage flicker.

[0088] For the specific steps of the method of the present invention, please refer to Figure 2 The flow chart shown.

Claims

1. A method for detecting a broken rotor bar fault of an asynchronous motor which can avoid the influence of voltage flicker, characterized in that: The asynchronous motor targeted by the fault detection method is a three-phase cage-type asynchronous motor. The detection device used comprises a voltage regulator (1), a current transformer (2), a voltage transformer (3), a data acquisition device (4), a computer (5), a load (6) and a tachometer (7), wherein the voltage regulator (1) is used to adjust the test voltage of the asynchronous motor D under test, the current transformer (2) and the voltage transformer (3) are used to collect the feeding current and voltage, the data acquisition device (4) sends the measured data to the computer (5) for calculation and analysis, the tachometer (7) is used to measure the rotation speed of the asynchronous motor, and the load (6) is connected to the output shaft of the asynchronous motor; During detection, when the power supply voltage flicker exists and affects the detection of the rotor bar fault, the stator current and voltage of the asynchronous motor are measured and analyzed, the main characteristic components of the rotor bar fault are examined, the main characteristic frequency, auxiliary characteristic frequency and power supply voltage flicker characteristic frequency of the rotor bar fault are calculated and compared, the characteristic frequency and component unique to the rotor bar fault are selected, and then the characteristic components are analyzed to accurately determine whether the rotor bar fault occurs; The main characteristic frequency and auxiliary characteristic frequency of the rotor broken bar fault are calculated according to the following formula: ； Calculate the characteristic frequency of the power supply voltage flicker close to the main and auxiliary characteristic frequencies of the rotor bar broken fault, and calculate it according to the following formula: ； In the formula, f1 represents the fundamental frequency; s represents the slip rate; m = k ( Z 2 / P )±1, where k = 0, 1, 2, …; Z 2 indicates the number of rotor slots; ν = 6 c +1, among which c = 0, ±1, ±2, ….

2. The method for detecting a broken rotor bar fault of an asynchronous motor that can avoid the influence of voltage flicker according to claim 1 is characterized in that: The specific detection and analysis are carried out according to the following steps: a. A rotor bar fault detection method based on stator current signal analysis is used to measure the stator current signal of one phase, and an adaptive filtering method is used to offset the stator current frequency component. A refined FFT spectrum analysis is performed to obtain a refined FFT spectrum. The presence of the main characteristic components of the rotor bar fault is examined to predict whether the rotor state is faulty. If a spectrum peak corresponding to the main characteristic components of the rotor bar fault appears in the refined FFT spectrum, the rotor bar is broken. If the spectrum peak does not exist in the refined FFT spectrum, the rotor is healthy. b. Detect power supply voltage flicker and determine whether it affects rotor bar fault detection First, calculate the single-phase supply voltage u according to f1 determined in step a. s ( t ) refines the FFT spectrum at f1 and determines the amplitude of its fundamental component; then, s (t) Adaptive filtering is performed to filter out the fundamental component, and a detailed FFT spectrum analysis is performed to determine the amplitude and frequency f of the power supply voltage flicker component. m ; Then, according to f m Is it close to 2 sf1? Determine whether the power supply voltage flicker affects the rotor broken bar fault detection. If f m If it is not close to 2 sf1, the power supply voltage flicker does not affect the rotor bar fault detection, the prediction of the rotor state in step a is established, and the rotor bar fault detection is completed; otherwise, the subsequent steps need to be further performed; c. Determine the characteristic frequency of the current rotor bar failure, which is unique to the current rotor bar failure and different from the supply voltage flicker, and record it as f RF ; First, the main and auxiliary characteristic frequencies of the broken rotor bar fault are calculated, and then the power supply voltage flicker characteristic frequencies close to the main and auxiliary characteristic frequencies of the broken rotor bar fault are calculated; Calculate the power supply voltage flicker characteristic frequency close to the rotor bar fault characteristic frequency, "close" means that the power supply voltage flicker characteristic frequency is within the range of [- 5Hz — + 5Hz] of the rotor bar fault characteristic frequency; Then, the characteristic frequency of the broken rotor bar fault is compared with the characteristic frequency of the power supply voltage flicker, so as to select the characteristic frequency f unique to the broken rotor bar fault. RF ; d. Check the refined FFT spectrum of the one-phase stator current signal obtained in step a. RF Whether there is a corresponding spectrum peak nearby, and make further accurate judgment: if the spectrum graph is at f RF If there is no spectrum peak near f, the rotor is healthy and only the supply voltage flickers. RF If a spectrum peak appears near the rotor bar, it means that the rotor bar is broken, accompanied by power supply voltage flicker.

Images