Outer rotor permanent magnet synchronous generator winding turn-to-turn short circuit fault diagnosis device and method

Abstract

The invention relates to the technical field of generator fault detection, in particular to an outer rotor permanent magnet synchronous generator winding turn-to-turn short circuit fault diagnosis device and method, the outer rotor permanent magnet synchronous generator winding turn-to-turn short circuit fault diagnosis device comprises a signal acquisition unit, a signal transmission unit and a signal processing unit, and the signal acquisition unit comprises a magnetic field detection module and a current detection module. The structure is reasonable and compact, when the rotor rotates, the detection coil induces the circumferential distribution change of an air-gap magnetic field in real time, detection signals of the detection coil are input into the outer copper ring through the wire, annular current is formed in the outer copper ring, and the annular current excites alternating magnetic fields distributed in an axial symmetry mode in the space around the outer copper ring. The static concentric winding induces an induced current in the concentric winding in a rotating alternating magnetic field, the current detection module obtains a fundamental current and a fault characteristic harmonic component, and early fault detection and short-circuit phase identification are realized by analyzing asymmetric harmonic characteristics in the induced current of the detection coil. The problem that a traditional diagnosis method fails in the severe environment of the petroleum industry is solved.

Description

Technical Field

[0001] This invention relates to the field of generator fault detection technology, and is a device and method for diagnosing inter-turn short-circuit faults in the windings of an external rotor permanent magnet synchronous generator. Background Technology

[0002] External rotor permanent magnet synchronous generators (PMSGs) are increasingly widely used in the petroleum industry due to their advantages of small size, high efficiency, and high power density, playing a particularly important role in oil drilling rig top drive systems, electric pumping units, offshore platform power generation systems, and oilfield distributed energy systems. However, the harsh conditions of the oilfield operating environment, including high temperature, high pressure, strong vibration, and oil corrosion, pose significant challenges to the operational reliability of external rotor PMSGs.

[0003] Stator winding inter-turn short circuit is one of the common early faults of external rotor permanent magnet synchronous generators. In oil drilling and production operations, due to the characteristics of frequent load fluctuations, frequent start-stops, and overload operation, the stator winding is subjected to complex electromagnetic and thermal stresses, making it extremely prone to insulation degradation. In the early stages of the fault, it can lead to localized heating and torque pulsation in the winding, eventually resulting in insulation breakdown and generator shutdown. For oil production, unplanned equipment shutdowns not only cause direct economic losses but may also trigger safety accidents such as well blowouts and environmental pollution, with extremely serious consequences.

[0004] Traditional detection methods include stator current analysis, temperature monitoring, and vibration signal analysis, but these methods all have significant limitations in the oilfield operating environment. (1) The stator current analysis method relies on the amplitude ratio of the fault current to the fundamental current to identify the fault. However, the load fluctuation of equipment such as oil drilling rigs and pumping units is severe, which will significantly interfere with the characteristic signal. Especially under the variable speed and load conditions during drilling, the diagnostic accuracy is low. (2) Temperature monitoring method requires the fault to develop to the point of local overheating before obvious signals appear. The response is delayed. In the high temperature environment of oil well site, the ambient temperature fluctuates greatly, making it difficult to accurately capture the temperature rise signal in the early insulation deterioration stage. (3) Vibration signal analysis is easily affected by mechanical noise such as drilling rig vibration, mud pump pulsation, and tubing resonance at the oil drilling site. In a strong electromagnetic environment, the signal-to-noise ratio is low and the fault location error is large.

[0005] Inter-turn short circuits in the stator windings can lead to an imbalance in the spatial distribution of the stator magnetomotive force, causing distortion of the air gap magnetic field and magnetic flux imbalance. Therefore, a detection coil can be installed on the stator side to check the magnetic flux balance, thereby identifying inter-turn short circuit faults. However, the structure of an external rotor permanent magnet synchronous generator is such that the rotor envelops the stator, making it difficult for traditional stator-side detection coils to fully cover the air gap magnetic field. Furthermore, existing contact-type signal transmission schemes often use slip rings. In the oily, muddy, and high-humidity environments of oil well sites, long-term operation of slip rings can cause severe mechanical wear and corrosion, easily leading to changes in contact resistance and signal distortion, resulting in high maintenance costs and short service life.

[0006] Meanwhile, oil drilling and production equipment often operates at variable speeds. Drilling rigs adjust their rotational speed based on formation hardness, and pumping units optimize stroke frequency based on fluid production. Fixed-frequency sampling leads to harmonic distortion of fault characteristics, further reducing diagnostic reliability. Currently, the industry lacks a systematic diagnostic solution for external rotor structures that can adapt to the variable speed and load conditions of the oil industry and withstand harsh environments. There is an urgent need for a non-contact, interference-resistant, and highly sensitive fault diagnosis method to ensure the continuity and safety of oil production. Summary of the Invention

[0007] This invention provides a device and method for diagnosing inter-turn short circuit faults in the windings of an external rotor permanent magnet synchronous generator, which overcomes the shortcomings of the prior art and can effectively solve the problem that traditional fault diagnosis methods for generators are prone to failure in the harsh environment of the petroleum industry.

[0008] One of the technical solutions of the present invention is achieved through the following measures: a short-circuit fault diagnosis device for inter-turn windings of an external rotor permanent magnet synchronous generator, comprising a signal acquisition unit, a signal transmission unit, and a signal processing unit; The signal acquisition unit includes a magnetic field detection module and a current detection module. The magnetic field detection module includes several detection coils distributed circumferentially on the outside of the rotor core of the outer rotor permanent magnet synchronous generator. The detection coils are connected in series, and each detection coil has a protective layer fixed on its outside. The signal transmission unit includes a rotating side assembly and a stationary side assembly. The stationary side assembly includes a concentric winding and a support. A fixed shaft is fixedly installed inside the support. The rear end of the fixed shaft is fixedly installed together with the front end of the stator of the external rotor permanent magnet synchronous generator. The concentric winding is spirally wound on the outer side of the fixed shaft. The rotating side assembly includes an outer copper ring and a connecting frame. The front end of the rotor of the outer rotor permanent magnet synchronous generator is fixedly installed with a connecting frame that is fitted on the outside of the fixed shaft. The inner side of the connecting frame is provided with an annular groove, and the outer copper ring is installed in the annular groove. There is a gap between the inner side of the outer copper ring and the outer side of the concentric winding. The current detection module includes a current transformer, with an outer copper ring connected to the detection coil. The current transformer is used to collect the induced current signal of the concentric winding. The signal processing unit includes an acquisition module, a processing module, and a host computer. The concentric winding and current transformer are connected to the acquisition module, the acquisition module is connected to the processing module, and the processing module is connected to the host computer.

[0009] The following are further optimizations and / or improvements to one of the above-mentioned technical solutions: The aforementioned magnetic field detection module may include three detection coils distributed circumferentially on the outside of the rotor core of the external rotor permanent magnet synchronous generator. The detection coils are parallel to the central axis of the external rotor permanent magnet synchronous generator, and the protective layer includes epoxy resin encapsulated on the outside of the detection coils.

[0010] The gap between the outer side of the concentric winding and the inner side of the outer copper ring can be 0.2 to 0.5 mm, and the coaxiality between the concentric winding and the outer copper ring is no more than 0.05 mm.

[0011] The thickness of the outer copper ring can be 3 to 5 mm, the surface roughness of the outer copper ring is not greater than 0.4 μm, and there is an interference fit between the outer copper ring and the ring groove, with an interference amount of 0.02 to 0.04 mm.

[0012] The second technical solution of the present invention is achieved through the following measures: a method for diagnosing inter-turn short-circuit faults in the windings of an external rotor permanent magnet synchronous generator, comprising the following steps: S1, the induced voltage signal generated by the change of the air gap magnetic field of the external rotor permanent magnet synchronous generator is input into the outer copper ring to form a ring current; S2, collects the induced current signal formed when the concentric winding rotates in the outer copper ring with a ring current; S3 employs a composite variational mode decomposition algorithm, using a two-level decomposition strategy to denoise and enhance the acquired induced current signal. S4. Obtain the current harmonic amplitude based on the induced current signal after noise reduction and signal enhancement processing, and obtain the fault index based on the current harmonic amplitude. S5 determines whether a short circuit fault has occurred between the turns of the stator winding of the external rotor permanent magnet synchronous generator based on the change range of the fault indicators.

[0013] The following are further optimizations and / or improvements to the second technical solution of the above invention: Step S3 above employs a composite variational mode decomposition algorithm, using a two-level decomposition strategy to denoise and enhance the acquired induced current signal, including: S31, set the induced current signal collected by the current transformer as the original signal f(t), and use the original signal f(t) as input to perform a CVMD decomposition. S32, Calculate the correlation coefficient between each IMF component and the original signal f(t). And set a discrimination threshold ε, when the correlation coefficient When the correlation coefficient is less than ε, the IMF component is dominated by noise. When ≥ε, the IMF component is an effective component; Where N is the signal length, the horizontal line represents the mean, and IMFi(n) is the instantaneous amplitude of the i-th intrinsic mode function at the n-th sampling point. Let f(n) be the average value of the i-th intrinsic mode function over all sampling points, and let f(n) be the amplitude of the original induced current signal at the n-th sampling point. This is the average value of the original induced current signal across all sampling points; S33, after removing the noise-dominated IMF components, the remaining IMF components are superimposed and reconstructed to obtain the initial denoised signal. S34, perform secondary CVMD decomposition on the initial denoised signal to obtain the induced current signal after denoising and signal enhancement processing.

[0014] In step S4 above, the fault indicator obtained based on the current harmonic amplitude is FI, which is obtained by the following formula: Under normal operating conditions, f1 and f2 are the current harmonic amplitudes of 1+1 / p and 1-1 / p in the detection coil, respectively; under inter-turn short-circuit fault conditions, f1 and f2 are obtained according to the following formula: Among them, f s denoted as the fundamental frequency of the stator current of the external rotor permanent magnet synchronous generator, and p is the number of pole pairs of the external rotor permanent magnet synchronous generator.

[0015] In step S4 above, the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator is... s The extraction steps are as follows: When the stator speed change rate does not exceed the set value, the Hanning window is used as the window function, and the window time is set to include 4-6 complete fundamental frequency cycles. The window function slides with a set step size. Within each time window, the windowed signal is subjected to FFT transformation, and the spectral line with the largest amplitude is searched in the range of 45-55Hz. This frequency is the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator at the current moment. s ; When the stator speed change rate exceeds the set value, the complex Morlet wavelet is used as the mother wavelet, and wavelet transform is performed within the scale range corresponding to 40-60Hz. The instantaneous frequency corresponding to the ridge with the largest modulus on the scale-time plane is the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator at the current moment. s .

[0016] Step S5 above determines whether a short-circuit fault has occurred between turns of the stator winding of the external rotor permanent magnet synchronous generator based on the change range of the fault index. Specifically, if the fault index FI increases to k under normal operating conditions... f If the value is multiple, then an inter-turn short circuit fault has occurred in the winding of the external rotor permanent magnet synchronous generator; otherwise, no inter-turn short circuit fault has occurred in the winding of the external rotor permanent magnet synchronous generator. f This is the short-circuit fault threshold.

[0017] This invention features a reasonable and compact structure. In use, three detection coils are evenly distributed along the outer circumference of the rotor core of the external rotor permanent magnet synchronous generator. The axis of each detection coil is parallel to the central axis of the stator. The inner surface of the detection coil is tightly fitted to the outer side of the rotor core to ensure maximum magnetic flux coupling coefficient. Several detection coils are connected in series, and both ends of the series connection are symmetrically fixed to the outer side of the outer copper ring by wires to form a differential detection structure, which can effectively suppress common-mode interference. The detection coils are wound with high-temperature resistant enameled wire with a temperature rating of H and a temperature resistance of 180℃. When the rotor of the external rotor permanent magnet synchronous generator rotates, the detection coils sense the circumferential distribution change of the air gap magnetic field in real time.

[0018] The current detection module uses a high-precision Hall current transformer to collect the three-phase stator current signal of the external rotor permanent magnet synchronous generator. The current transformer can simultaneously acquire the fundamental current and the fault characteristic harmonic components, providing auxiliary criteria for fault diagnosis.

[0019] The detection signal from the detection coil is input into the outer copper ring through the wire, forming a circular current inside the outer copper ring. According to Ampere's circuital law, the circular current excites an axisymmetrically distributed alternating magnetic field in the space around the outer copper ring. The stationary concentric winding is in the rotating alternating magnetic field, and according to Faraday's law of electromagnetic induction, an induced current is induced in the concentric winding.

[0020] The acquisition module can dynamically adjust the sampling frequency according to the real-time rotor speed to ensure that a fixed number of points are collected in each mechanical rotation cycle. This "equal angle sampling" strategy fundamentally eliminates the distortion effect of speed fluctuation on the signal spectrum, providing a stable and reliable digital signal source for the subsequent accurate extraction of fault feature harmonics. It overcomes the problem of fault feature frequency shift caused by the variable speed operation of generators in the oil field. The processing module transmits the processed signal to the host computer for in-depth analysis and diagnosis of fault features.

[0021] This application enables non-contact signal transmission, anti-interference detection, and variable speed adaptive diagnosis. By analyzing the asymmetric harmonic characteristics in the induced current of the detection coil, it achieves early fault detection, short circuit phase identification, and quantitative assessment of severity, ensuring the continuity and safety of oil production and solving the problem of the failure of traditional diagnostic methods in the harsh environment of the oil industry. Attached Figure Description

[0022] Appendix Figure 1 These are schematic diagrams of the structures of embodiments one to four of the present invention.

[0023] Appendix Figure 2 This is a partial cross-sectional view of the fixed shaft in embodiments one to four of the present invention.

[0024] Appendix Figure 3 This is a side view schematic diagram of the detection coil in embodiments one to four of the present invention.

[0025] Appendix Figure 4 This is a graph showing the induced current curve of the winding of an external rotor permanent magnet synchronous generator under normal operating conditions.

[0026] Appendix Figure 5 The induced current curve is a simulation of an inter-turn short-circuit fault in the winding of an external rotor permanent magnet synchronous generator.

[0027] Appendix Figure 6 The spectrum of induced current during a simulated inter-turn short-circuit fault in the winding of an external rotor permanent magnet synchronous generator.

[0028] The codes in the attached diagram are as follows: 1 is the outer rotor, 2 is the detection coil, 3 is the concentric winding, 4 is the support, 5 is the fixed shaft, 6 is the outer copper ring, 7 is the connecting frame, 8 is the current transformer, 9 is the acquisition module, 10 is the processing module, and 11 is the host computer. Detailed Implementation

[0029] The present invention is not limited to the following embodiments, and the specific implementation can be determined according to the technical solution of the present invention and the actual situation.

[0030] In this invention, for ease of description, the description of the relative positions of the components is based on the appendix to the specification. Figure 1 The layout is described using a diagrammatic method, such as the positional relationships of front, back, top, bottom, left, and right, which are based on the instructions attached. Figure 1 The orientation of the layout is determined by the direction of the map.

[0031] The present invention will be further described below with reference to embodiments and accompanying drawings: Example 1: As shown in the attached document Figure 1 , 2 As shown, the external rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnosis device includes a signal acquisition unit, a signal transmission unit, and a signal processing unit; The signal acquisition unit includes a magnetic field detection module and a current detection module. The magnetic field detection module includes several detection coils 2 distributed circumferentially on the outside of the rotor core of the outer rotor permanent magnet synchronous generator. The several detection coils 2 are connected in series, and each detection coil 2 has a protective layer fixed on its outside. The signal transmission unit includes a rotating side assembly and a stationary side assembly. The stationary side assembly includes a concentric winding 3 and a support 4. A fixed shaft 5 is fixedly installed inside the support 4. The rear end of the fixed shaft 5 is fixedly installed together with the front end of the stator of the external rotor permanent magnet synchronous generator. The concentric winding 3 is spirally wound on the outer side of the fixed shaft 5. The rotating side assembly includes an outer copper ring 6 and a connecting frame 7. The connecting frame 7, which is fitted on the outside of the fixed shaft 5, is fixedly installed at the front end of the rotor of the outer rotor permanent magnet synchronous generator. The connecting frame 7 has an annular groove on its inner side, and the outer copper ring 6 is installed in the annular groove. There is a gap between the inner side of the outer copper ring 6 and the outer side of the concentric winding 3. The current detection module includes a current transformer 8, an outer copper ring 6 connected to the detection coil 2, and the current transformer 8 is used to collect the induced current signal of the concentric winding 3. The signal processing unit includes an acquisition module 9, a processing module 10, and a host computer 11. The concentric winding 3 and the current transformer 8 are connected to the acquisition module 9, the acquisition module 9 is connected to the processing module 10, and the processing module 10 is connected to the host computer 11.

[0032] In use, three detection coils 2 are evenly distributed (with a 120° interval between adjacent detection coils 2) along the outer circumference of the rotor core of the external rotor permanent magnet synchronous generator. The axis of each detection coil 2 is parallel to the central axis of the stator. The inner surface of the detection coil 2 is in close contact with the outer side of the rotor core to ensure maximum magnetic flux coupling coefficient. The three detection coils 2 are connected in series. The two ends of the series connection of the three detection coils 2 are symmetrically fixed to the outer side of the outer copper ring 6 by wires to form a differential detection structure, which can effectively suppress common mode interference. The detection coils 2 are wound with high temperature resistant enameled wire with a temperature rating of H and a temperature resistance of 180℃. When the rotor of the external rotor permanent magnet synchronous generator rotates, the detection coils 2 sense the circumferential distribution change of the air gap magnetic field in real time.

[0033] One of the two connecting wires between the two ends of the concentric winding 3 and the acquisition module 9 passes through the current transformer 8. The current detection module uses a high-precision Hall current transformer 8 (accuracy class 0.2, frequency response range DC-10kHz) to acquire the three-phase stator current signal of the external rotor permanent magnet synchronous generator. The current transformer 8 can simultaneously acquire the fundamental current and the fault characteristic harmonic components, providing auxiliary criteria for fault diagnosis.

[0034] The electromagnetic coupling non-contact transmission structure formed by the rotating side component and the stationary side component of the signal transmission unit, the concentric winding 3 adopts QZY-2 / 180 grade oil-resistant polyesterimide enameled copper wire, and is precisely wound into a loop coil in a concentric spiral manner. The support 4 is set in front of the rotor permanent magnet synchronous generator and is made of non-magnetic stainless steel.

[0035] The detection signal of the detection coil 2 is input into the outer copper ring 6 through the wire, forming a circular current in the outer copper ring 6. According to Ampere's circuital law, the circular current excites an axisymmetrically distributed alternating magnetic field in the space around the outer copper ring 6. The stationary concentric winding 3 is in the rotating alternating magnetic field, and according to Faraday's law of electromagnetic induction, an induced current is induced in the concentric winding 3.

[0036] The acquisition module 9 is a known acquisition card, such as a USB series multi-functional current acquisition card. The acquisition module 9 has a frequency conversion sampling function. The acquisition module 9 can dynamically adjust the sampling frequency according to the real-time rotor speed to ensure that a fixed number of points (N points) are collected in each mechanical rotation cycle. This "equal angle sampling" strategy fundamentally eliminates the distortion effect of speed fluctuation on the signal spectrum, and provides a stable and reliable digital signal source for the subsequent accurate extraction of fault feature harmonics. It overcomes the problem of fault feature frequency offset caused by the variable speed operation of the generator in the oil field. The processing module 10 is a known technology, such as a PLC. The processing module 10 transmits the processed signal to the host computer 11 for in-depth analysis and diagnosis of fault features.

[0037] The above-mentioned external rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnosis device can be further optimized and / or improved according to actual needs: Example 2: As an optimization of the above examples, as shown in the appendix. Figure 1 , 3 As shown, the magnetic field detection module includes three detection coils 2 distributed circumferentially on the outside of the rotor core of the outer rotor permanent magnet synchronous generator. The detection coils 2 are parallel to the central axis of the outer rotor permanent magnet synchronous generator, and the protective layer includes epoxy resin encapsulated on the outside of the detection coils 2.

[0038] The outer layer of the detection coil 2 is encapsulated and fixed with oil-resistant and corrosion-resistant modified epoxy resin (not shown in the figure) to meet the long-term stable operation requirements of oil well sites in high-temperature (≥150℃) and oily environments.

[0039] Example 3: As an optimization of the above examples, as shown in the appendix. Figure 1 , 2 As shown, the gap between the outer side of the concentric winding 3 and the inner side of the outer copper ring 6 is 0.2 to 0.5 mm, and the coaxiality between the concentric winding 3 and the outer copper ring 6 is no greater than 0.05 mm.

[0040] A uniform air gap of 0.2-0.5mm is maintained between the outer side of the concentric winding 3 and the inner side of the outer copper ring 6 (multiple non-magnetic pads can be evenly distributed along the circumference between the outer side of the concentric winding 3 and the inner side of the outer copper ring 6 for positioning). This can form an efficient electromagnetic coupling channel. The support 4 supports the concentric winding 3 through the bearing to ensure that the coaxiality deviation with the outer copper ring 6 is ≤0.05mm.

[0041] Example 4: As an optimization of the above examples, as shown in the appendix. Figure 1 , 2 As shown, the thickness of the outer copper ring 6 is 3 to 5 mm, the surface roughness of the outer copper ring 6 is no greater than 0.4 μm, and the outer copper ring 6 and the ring groove are interference fit with an interference amount of 0.02 to 0.04 mm.

[0042] The outer copper ring 6 is made of oxygen-free copper material with high conductivity. The inner wall surface is precision machined and electrolytically polished, with a surface roughness of no more than 0.4μm, which can significantly reduce eddy current loss. The outer copper ring 6 has a wall thickness of 3 to 5mm, which ensures both structural strength and maintains an appropriate resistance value to form a stable ring current.

[0043] The connecting frame 7 is made of high-strength aluminum alloy material, and the inner surface is precision machined with an annular groove (tolerance grade IT7). The outer copper ring 6 is embedded in the groove with an interference fit, and the interference is 0.02 to 0.04 mm. This can ensure reliable fixation of the outer rotor 1 of the outer rotor permanent magnet synchronous generator when it rotates at high speed.

[0044] The outer rotor of the permanent magnet synchronous generator has four pre-drilled M8 threaded holes on the front side of the outer rotor 1 end cover, which are evenly distributed at 90°. The end cover and the connecting frame 7 are connected by stainless steel double-ended studs, and anti-loosening nuts and spring washers are used to ensure the reliability of the connection under vibration environment.

[0045] Example 5: As attached Figure 1 As shown, the method for diagnosing inter-turn short-circuit faults in the windings of this external rotor permanent magnet synchronous generator includes the following steps: S1, the induced voltage signal generated by the change of the air gap magnetic field of the external rotor permanent magnet synchronous generator is input into the outer copper ring 6 to form a ring current; S2, collect the induced current signal formed when the concentric winding 3 rotates in the outer copper ring 6 with an annular current; S3 employs a composite variational mode decomposition algorithm, using a two-level decomposition strategy to denoise and enhance the acquired induced current signal. S4. Obtain the current harmonic amplitude based on the induced current signal after noise reduction and signal enhancement processing, and obtain the fault index based on the current harmonic amplitude. S5 determines whether a short circuit fault has occurred between the turns of the stator winding of the external rotor permanent magnet synchronous generator based on the change range of the fault indicators.

[0046] The proposed method for diagnosing inter-turn short circuit faults in the windings of an external rotor permanent magnet synchronous generator is reliable and easy to implement. It avoids the mechanical wear and signal interference associated with traditional contact transmission and can solve the problem of detecting inter-turn short circuits in the stator windings of an external rotor permanent magnet synchronous generator when it is rotating using a small-volume, low-cost, and practical device.

[0047] This application enables non-contact signal transmission, anti-interference detection, and variable speed adaptive diagnosis. By analyzing the asymmetric harmonic characteristics in the induced voltage of the detection coil 2, it achieves early fault detection, short circuit phase identification, and severity quantification assessment, ensuring the continuity and safety of oil production and solving the technical problem of the failure of traditional diagnostic methods in the harsh environment of the oil industry.

[0048] The above-mentioned method for diagnosing inter-turn short-circuit faults in the windings of external rotor permanent magnet synchronous generators can be further optimized and / or improved according to actual needs: Example 5: As an optimization of the above example, step S3 employs a composite variational mode decomposition algorithm to perform denoising and signal enhancement processing on the acquired induced current signal through a two-level decomposition strategy, including: S31, set the induced current signal collected by the current transformer 8 as the original signal f(t), and use the original signal f(t) as input to perform a CVMD decomposition. S32, Calculate the correlation coefficient between each IMF component and the original signal f(t). And set a discrimination threshold ε, when the correlation coefficient When the correlation coefficient is less than ε, the IMF component is dominated by noise. When ≥ε, the IMF component is an effective component; Where N is the signal length, the horizontal line represents the mean, and IMFi(n) is the instantaneous amplitude of the i-th intrinsic mode function at the n-th sampling point. Let f(n) be the average value of the i-th intrinsic mode function over all sampling points, and let f(n) be the amplitude of the original induced current signal at the n-th sampling point. This is the average value of the original induced current signal across all sampling points; S33, after removing the noise-dominated IMF components, the remaining IMF components are superimposed and reconstructed to obtain the initial denoised signal. S34, perform secondary CVMD decomposition on the initial denoised signal to obtain the induced current signal after denoising and signal enhancement processing.

[0049] A composite variational mode decomposition algorithm is adopted, and a two-level decomposition strategy is used to denoise and enhance the acquired induced current signal.

[0050] In the complex electromagnetic environment of oil well sites, the induced voltage signal collected by probe coil 2 is inevitably subject to various interferences: high-frequency switching noise generated by the frequency converter, electromagnetic radiation from nearby equipment, baseline drift caused by mechanical vibration, and random white noise, etc. This invention employs a composite variational mode decomposition (CVMD) algorithm, using a two-level decomposition strategy to achieve noise suppression and signal enhancement, thus avoiding severe interference that could significantly affect the accuracy of fault feature extraction.

[0051] The first CVMD decomposition adopts an adaptive mode number determination strategy: the induced current signal is set as the original signal f(t), and the original signal f(t) is used as input to perform a CVMD decomposition. Specifically, the decomposition mode number is incremented starting from k1=2, and the center frequency change of each intrinsic mode function (IMF) is monitored. When the center frequency change rate of two adjacent decompositions is less than 1% and the frequency bands of each IMF component do not overlap, the optimal mode number is determined.

[0052] The decomposed IMF components have different physical meanings: the high-frequency IMF components (usually IMF1-IMF3) mainly contain broadband noise and switching harmonics; the mid-frequency IMF components (IMF4-IMF5) contain fault characteristic information; and the low-frequency IMF components (IMF6-IMF7) correspond to the fundamental component and low-frequency disturbances.

[0053] To accurately identify the effective components, the correlation coefficient between each IMF component and the original signal f(t) is calculated. And set the discrimination threshold ε=0.1, when When <ε, the IMF component is dominated by noise. When ≥ε, the IMF component is an effective component; Where N is the signal length, the horizontal line represents the mean, and IMFi(n) is the instantaneous amplitude of the i-th intrinsic mode function at the n-th sampling point. Let f(n) be the average value of the i-th intrinsic mode function over all sampling points, and let f(n) be the amplitude of the original induced current signal at the n-th sampling point. This is the average value of the original induced current signal across all sampling points. This threshold is derived based on a large amount of field data and achieves the best balance between noise suppression and signal fidelity. After removing the noise-dominated IMF components, the remaining IMF components are superimposed and reconstructed to obtain the initial denoised signal. The initial denoised signal is subjected to secondary CVMD decomposition to obtain the denoised and signal-enhanced induced current signal.

[0054] The number of modes k2 in the above-mentioned secondary CVMD decomposition can be set to 5 to accurately separate the fault characteristic frequency band. The rest of the decomposition process is the same as that of the primary CVMD decomposition, and will not be described in detail here.

[0055] Furthermore, the augmented Dickey-Fuller (ADF) test can be used to assess the stationarity of each IMF component. When the test statistic is less than the critical value of the 5% significance level, it is determined to be a stationary sequence. Non-stationary components are converted into stationary sequences by difference or trend term removal to ensure the effectiveness of subsequent spectral analysis. After two-stage CVMD processing, not only is noise effectively suppressed, but the non-stationary signal is also transformed into a combination of multiple stationary components, laying the foundation for accurate fault feature extraction.

[0056] Example 6: As an optimization of the above example, in step S4, the fault indicator FI is obtained based on the current harmonic amplitude. FI is obtained by the following formula: Under normal operating conditions, f1 and f2 are the current harmonic amplitudes of 1+1 / p and 1-1 / p in detection coil 2, respectively; under inter-turn short-circuit fault conditions, f1 and f2 are obtained according to the following formula: Among them, f s denoted as the fundamental frequency of the stator current of the external rotor permanent magnet synchronous generator, and p is the number of pole pairs of the external rotor 1 permanent magnet generator.

[0057] In the rotor coordinate system, the air gap magnetic flux density distribution of the external rotor 1 permanent magnet generator under normal operating conditions is B. normal (θ) r ), B normal (θ) r It is obtained through the following formula: Among them, B m denoted as ρ, where p is the air gap magnetic flux density amplitude of the external rotor 1 permanent magnet generator under normal operating conditions. δ is the spatial position angle of the external rotor 1 permanent magnet generator in the rotor coordinate system, and δ is the phase angle of the external rotor 1 permanent magnet generator; The air gap magnetic flux density distribution of the external rotor 1 permanent magnet generator under an inter-turn short-circuit fault is B. fault (θ) r ), B fault (θ) r It is obtained through the following formula: Where Bs is the air gap magnetic flux density amplitude of the external rotor 1 permanent magnet generator under inter-turn short circuit fault, n is a positive integer, ω is the electrical angle of the external rotor 1 permanent magnet generator, t is time, δs is the phase angle, and μ is the time order of the magnetic field under normal operating conditions, taking an odd number. The induced current of a single detection coil 2 is i, which is obtained by the following formula: Where S is the area of ​​detection coil 2, and λ represents the spatial position angles corresponding to the two sides of the detection coil 2, λ is the proportionality coefficient between the leakage magnetic field and the air gap magnetic field, which is a constant less than 1, e is the voltage induced by the detection coil 2, and R is the resistance of the circuit of the detection coil 2. Substituting the air gap magnetic flux density distribution of the external rotor 1 permanent magnet generator under normal operating conditions and inter-turn short circuit fault respectively, we can obtain the induced current of a single detection coil 2 under normal operating conditions: i=0; And the induced current of a single detection coil 2 under an inter-turn short-circuit fault: Where S is the area of ​​detection coil 2, and denoted as the spatial position angles corresponding to the two sides of the detection coil 2, λ is the proportionality coefficient between the leakage magnetic field and the air gap magnetic field, which is a constant less than 1, e is the voltage induced by the detection coil 2, and R is the resistance of the circuit of the detection coil 2.

[0058] When the external rotor 1 permanent magnet generator is running normally, the air gap magnetic field is circumferentially symmetrically distributed in the rotor coordinate system; when a short circuit occurs between turns of the stator winding, the magnetic field symmetry is destroyed, generating characteristic harmonic components.

[0059] Stator winding inter-turn short circuits typically occur only within a single slot (i.e., under a specific pole pair), and the probability of multiple simultaneous short circuits in actual operation is extremely low, therefore it is not considered. Based on the local impact range of the fault, the generator space can be divided into short-circuit pole region and non-short-circuit pole region. For example, for a p-pole generator, the air gap magnetic flux density amplitude under the short-circuit pole (p-1 pole pair) is Bs1, and under the non-short-circuit pole (p-1 pole pair) is Bs3. When the detection coil 2 is in the non-short-circuit pole region, its magnetic field remains unchanged, so it basically does not generate an induced current. When the detection coil 2 transitions from the non-short-circuit pole region to the short-circuit region, the air gap magnetic field changes, thus generating a more significant induced current. The formula for calculating the air gap magnetic flux density under inter-turn short-circuit faults is applicable to the magnetic fields of different regions.

[0060] The detection coil 2 is fixed to the outer surface of the rotor core and is stationary relative to the rotor core.

[0061] Under normal operating conditions, the detection coil 2 is in a constant magnetic field and the magnetic flux density does not change, so no induced current is generated. After a short circuit fault, due to the disruption of the circumferential symmetry of the magnetic field, fractional harmonics such as 1±1 / p, 1±2 / p, 3±1 / p, and 3±2 / p will be generated, which are the sideband components of odd-order harmonics. For k detection coils 2 connected in series, the current of the detection coil 2 can be regarded as k times that of a single coil.

[0062] Furthermore, taking a 12-pole / 54-slot external rotor permanent magnet synchronous generator as the object, the current of the detection coil 2 under normal and short-circuit fault conditions was calculated in ANSYS Workbench software. The simulation results are as follows: Figures 4 to 6 As shown.

[0063] The calculation results show that the induced current in detection coil 2 is basically zero under normal operating conditions. However, due to factors such as approximate model boundary conditions and mesh generation accuracy, a very small induced current exists in detection coil 2. Figure 4 As shown. When a short-circuit fault occurs, the amplitude of the current in detection coil 2 increases sharply to 0.6mA. Furthermore, Fourier decomposition of the time-domain signal reveals that the spectrum of the current in detection coil 2 is dominated by fractional harmonics such as 1±1 / p, 1±2 / p, and 1±3 / p. Figure 5 , 6 As shown. Therefore, by comparing the difference in induced current under normal operating conditions and inter-turn short-circuit faults, inter-turn short-circuit faults in the stator winding of the external rotor 1 permanent magnet generator can be accurately identified.

[0064] Example 6: As an optimization of the above example, in step S4, the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator is... s The extraction steps are as follows: When the stator speed change rate does not exceed the set value, the Hanning window is used as the window function, and the window time is set to include 4-6 complete fundamental frequency cycles. The window function slides with a set step size. Within each time window, the windowed signal is subjected to FFT transformation, and the spectral line with the largest amplitude is searched in the range of 45-55Hz. This frequency is the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator at the current moment. s ; When the stator speed change rate exceeds the set value, the complex Morlet wavelet is used as the mother wavelet, and wavelet transform is performed within the scale range corresponding to 40-60Hz. The instantaneous frequency corresponding to the ridge with the largest modulus on the scale-time plane is the fundamental frequency f of the stator current of the external rotor permanent magnet synchronous generator at the current moment. s .

[0065] In oil drilling operations, generator speed varies frequently with drilling depth, formation hardness, and process requirements, making it difficult for traditional fixed-frequency analysis methods to accurately extract fault characteristics. This application uses a current transformer 8 to collect stator phase current signals and employs time-frequency analysis technology to achieve speed tracking and adaptive fault diagnosis.

[0066] Specifically, the three-phase current signals ia(t), ib(t), and ic(t) are first obtained by the current transformer 8 installed at the stator output terminal. A Hall-type current transformer 8 with an accuracy of 0.2 class and a bandwidth of DC-10kHz is selected to ensure accurate capture of the fundamental and harmonic components. After the signal is sampled by the anti-aliasing filter and the 16-bit ADC acquisition module 9, it is sent to the signal processing module 10 at a sampling rate of 10kS / s.

[0067] For real-time extraction of the fundamental frequency, this application employs Short-Time Fourier Transform (STFT) as the primary method. The setpoint is 5% / second. When the stator speed change rate does not exceed 5% / second, a Hanning window is selected as the window function, with a window length set to 0.1 seconds to ensure 4-6 complete fundamental cycles are included. This guarantees a frequency resolution of 0.01Hz while providing rapid response to speed changes. The window function slides in 0.01-second steps to achieve a 100Hz update rate. Within each time window, an FFT is performed on the windowed signal, searching for the spectral line with the largest amplitude within the 45-55Hz range. This frequency is the fundamental frequency f of the stator current at the current moment. s To improve the accuracy of frequency estimation, a three-spectral-line interpolation algorithm is adopted. By using the amplitude of the peak spectral line and its two adjacent spectral lines, a precise frequency value is obtained through quadratic curve fitting, which improves the frequency resolution by more than 10 times.

[0068] When a rapid change in rotational speed is detected (the rate of change exceeds 5% / second), the system automatically switches to Continuous Wavelet Transform (CWT) mode. The complex Morlet wavelet is used as the mother wavelet, whose excellent time-frequency localization characteristics are particularly suitable for analyzing non-stationary signals. By adjusting the scaling parameter, wavelet transform is performed within a scale range corresponding to 40-60Hz. The instantaneous frequency corresponding to the ridge with the largest modulus on the scale-time plane is the fundamental frequency f of the stator current. s .

[0069] Example 7: As an optimization of the above example, step S5 determines whether a short circuit fault has occurred between turns of the stator winding of the external rotor 1 permanent magnet generator based on the change range of the fault index. Specifically, if the fault index FI increases to k under normal operating conditions... f If the value is multiple, then an inter-turn short circuit fault has occurred in the winding of the external rotor permanent magnet synchronous generator; otherwise, no inter-turn short circuit fault has occurred in the winding of the external rotor permanent magnet synchronous generator. f This is the short-circuit fault threshold.

[0070] Under normal operating conditions, there is no induced current in detection coil 2. After a short-circuit fault, an induced current appears, containing harmonics of orders 1±1 / p and 1±2 / p. Due to certain errors and noise, some current also exists under normal operating conditions, but the harmonic components of the current under a short-circuit fault are still significant. Therefore, the occurrence of a short-circuit fault can be measured by calculating the fault index (harmonic amplitude). When FI increases to k under normal operating conditions... f When the value is multiple, an inter-turn short-circuit fault can be considered to have occurred. f The short-circuit fault threshold is set based on the motor's structural and material parameters.

[0071] The above technical features constitute various embodiments of the present invention, which have strong adaptability and implementation effect. Unnecessary technical features can be added or removed according to actual needs to meet the needs of different situations.

Claims

1. An outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnostic device characterized by The signal acquisition unit, the signal transmission unit and the signal processing unit are included. The signal acquisition unit includes a magnetic field detection module and a current detection module, the magnetic field detection module includes a plurality of detection coils which are distributed on the outer side of the rotor core of the outer rotor permanent magnet synchronous generator, the plurality of detection coils are connected in series, and each detection coil is fixed with a protective layer on the outer side thereof. The signal transmission unit includes a rotating side assembly and a stationary side assembly, the stationary side assembly includes a concentric winding and a support, a fixed shaft is fixedly installed in the support, the rear end of the fixed shaft is fixedly installed with the front end of the stator of the outer rotor permanent magnet synchronous generator, and the concentric winding is spirally arranged on the outer side of the fixed shaft. The rotating side assembly includes an outer copper ring and a connecting frame, the front end of the rotor of the outer rotor permanent magnet synchronous generator is fixedly installed with the connecting frame which is sleeved on the outer side of the fixed shaft, the inner side of the connecting frame is provided with a ring groove, and the outer copper ring is installed in the ring groove. The current detection module includes a current transformer, the outer copper ring is connected with the detection coil, and the current transformer is used for collecting the induced current signal of the concentric winding. The signal processing unit includes an acquisition module, a processing module and an upper computer, the concentric winding and the current transformer are connected with the acquisition module, the acquisition module is connected with the processing module, and the processing module is connected with the upper computer.

2. The outer rotor permanent magnet synchronous generator inter-turn short circuit fault diagnostic device according to claim 1, characterized by The magnetic field detection module includes three detection coils which are distributed on the outer side of the rotor core of the outer rotor permanent magnet synchronous generator, the detection coils are parallel to the central axis of the outer rotor permanent magnet synchronous generator, and the protective layer includes epoxy resin which is encapsulated on the outer side of the detection coil.

3. The outer rotor permanent magnet synchronous generator inter-turn short circuit fault diagnostic device according to claim 1 or 2, characterized by The gap between the outer side of the concentric winding and the inner side of the outer copper ring is 0.2-0.5 mm, and the coaxiality of the concentric winding and the outer copper ring is not greater than 0.05 mm.

4. The outer rotor permanent-magnet synchronous generator winding inter-turn short circuit fault diagnostic device according to claim 1 or 2, characterized by The thickness of the outer copper ring is 3-5 mm, the surface roughness of the outer copper ring is not greater than 0.4 μm, the outer copper ring is in interference fit with the ring groove, and the interference amount is 0.02-0.04 mm.

5. The outer rotor permanent-magnet synchronous generator winding inter-turn short circuit fault diagnostic device according to claim 3, characterized by The thickness of the outer copper ring is 3-5 mm, the surface roughness of the outer copper ring is not greater than 0.4 μm, the outer copper ring is in interference fit with the ring groove, and the interference amount is 0.02-0.04 mm.

6. An outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnosis method using the outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnosis device according to any one of claims 1 to 5, characterized by The steps are as follows: S1, inputting the induced voltage signal formed by the air gap magnetic field change of the outer rotor permanent magnet synchronous generator into the outer copper ring to form a ring current; S2, collecting the induced current signal formed when the concentric winding rotates in the outer copper ring with the ring current; S3, adopting a composite variational mode decomposition algorithm and a two-stage decomposition strategy to perform denoising and signal enhancement processing on the collected induced current signal; S4, obtaining the current harmonic amplitude according to the denoised and signal enhanced induced current signal, and obtaining the fault index according to the current harmonic amplitude; S5, judging whether the short circuit fault occurs between the turns of the stator winding of the outer rotor permanent magnet synchronous generator according to the change amplitude of the fault index.

7. The outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnostic method according to claim 6, characterized in that In step S3, the composite variational mode decomposition algorithm and the two-stage decomposition strategy are adopted to perform denoising and signal enhancement processing on the collected induced current signal, including: S31, setting the induced current signal collected by the current transformer as an original signal f(t), and performing once CVMD decomposition on the original signal f(t) as the input. S32, calculate the correlation coefficient of each IMF component and the original signal f(t) and set a discrimination threshold ε, when the correlation coefficient < ε, then the IMF component is dominated by noise, when the correlation coefficient ≥ ε, then the IMF component is an effective component; where N is the signal length, the horizontal line represents the mean value, IMFi(n) is the instantaneous amplitude of the i-th eigenmode function at the n-th sampling point, is the average value of the i-th eigenmode function over all sampling points, and f(n) is the amplitude of the original induction current signal at the n-th sampling point, is the average value of the original induction current signal over all sampling points. S33, after removing the IMF component dominated by noise, the remaining IMF components are superimposed to reconstruct a preliminary denoising signal; S34, the preliminary denoising signal is subjected to secondary CVMD decomposition to obtain an induced current signal after denoising and signal enhancement processing.

8. The outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnostic method according to claim 6 or 7, characterized in that In step S4, the fault index is obtained according to the current harmonic amplitude, and the fault index is FI, which is obtained by the following formula: Wherein, under normal working condition, f1 and f2 are current harmonic amplitudes of 1+1 / p and 1-1 / p in the detection coil respectively; under inter-turn short circuit fault, f1 and f2 are obtained according to the following formula: wherein f s is the fundamental frequency of the stator current of the outer rotor permanent magnet synchronous generator, and p is the number of pole pairs of the outer rotor permanent magnet synchronous generator.

9. The outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnostic method according to claim 8, characterized in that The stator current fundamental frequency f of the outer rotor permanent magnet synchronous generator in step S4 s The extraction step is as follows: when the stator speed change rate does not exceed a set value, a Hanning window is used as a window function, a window length time is set, the window length time contains 4-6 complete fundamental wave periods, the window function is slid with a set step, in each time window, the FFT transformation is performed on the windowed signal, and the spectral line with the maximum amplitude in the range of 45-55 Hz is searched, and the frequency is the stator current fundamental frequency f of the outer rotor permanent magnet synchronous generator at the current time s ; When the stator speed change rate exceeds the set value, the complex Morlet wavelet is taken as the mother wavelet, the wavelet transform is carried out in the scale range corresponding to 40-60Hz, and the instantaneous frequency corresponding to the ridge line with the maximum modulus value on the scale-time plane is the stator current fundamental frequency f of the outer rotor permanent magnet synchronous generator at the current time s .

10. The outer rotor permanent magnet synchronous generator winding inter-turn short circuit fault diagnostic method according to claim 9, characterized in that Step S5 judges whether the short circuit fault of the turn-to-turn of the stator winding of the outer rotor permanent magnet synchronous generator occurs according to the change range of the fault index, specifically: if the fault index FI increases to k f times of the normal working condition, the winding of the outer rotor permanent magnet synchronous generator occurs the turn-to-turn short circuit fault, otherwise, the winding of the outer rotor permanent magnet synchronous generator does not occur the turn-to-turn short circuit fault, k f is the short circuit fault threshold value.

Images