Method and device for detecting faults of wind turbine based on magnetic flux leakage detection

Abstract

The application relates to a wind driven generator fault detection method and device based on magnetic flux leakage detection, which comprises the following steps: obtaining an external magnetic flux leakage signal of a wind driven generator; adopting a variational mode decomposition method to decompose the external magnetic flux leakage signal into a plurality of intrinsic mode function signal components; determining the correlation of each intrinsic mode function signal component with the external magnetic flux leakage signal; selecting the first N intrinsic mode function signal components with relatively large correlation to reconstruct the external magnetic flux leakage signal and obtaining a reconstructed signal; and judging whether the wind driven generator has a fault according to the reconstructed signal. The wind driven generator fault detection method based on magnetic flux leakage detection can monitor the wind driven generator based on an external magnetic flux leakage array signal, adopts a signal reconstruction method, can preferably meet the requirements of non-intrusive, low cost and reliability, and can effectively extract and identify the stator and rotor inter-turn short circuit fault characteristics of the wind driven generator.

Description

Technical Field

[0001] This application relates to the field of wind turbine fault diagnosis technology, and more specifically, to a wind turbine fault detection method and device based on leakage magnetic field detection. Background Technology

[0002] With the rapid development of wind power technology and installed capacity, the cumulative operating time of wind turbines continues to increase, and the number of turbines nearing the end of their service life is constantly rising. The problems of wind turbine failure and maintenance are becoming increasingly prominent. As a key component of wind turbines, the health status of the wind turbine greatly affects the remaining lifespan of the unit. Aging of internal and external structural components increases the risk of generator failure. Currently, electrical fault monitoring of wind turbines is not yet a standard practice in the wind energy industry. Most condition monitoring systems and data acquisition and monitoring systems (SCADA) in the wind energy industry rely on monitoring high-frequency vibration signals from the wind turbine gearbox and bearings, generator current signals, or winding temperatures.

[0003] Given the relatively limited monitoring methods for wind turbines and the increasing focus on the reliability of key electrical components, improving the electrical fault monitoring capabilities of wind turbines by expanding the capabilities of the condition monitoring system without altering the existing structure of the existing units has practical engineering value.

[0004] External leakage flux is the magnetic flux that escapes from the air gap magnetic field of a wind turbine to the outside of the turbine casing. It is a projection of the air gap magnetic field from the inside out and contains state information of the wind turbine, which can be used to monitor the health status of the wind turbine. However, the external leakage flux signal of a wind turbine is generally very weak and is easily affected by environmental noise and interference signals. Current research mainly focuses on laboratory environments and does not consider the influence of real-world environmental noise and interference sources. Summary of the Invention

[0005] To overcome at least one deficiency in the prior art, this application provides a method and apparatus for wind turbine fault detection based on magnetic flux leakage detection.

[0006] Firstly, a method for fault detection of wind turbine generators based on magnetic flux leakage detection is provided, including:

[0007] Acquire the external leakage magnetic field signal of the wind turbine;

[0008] The variational mode decomposition method is used to decompose the external leakage magnetic signal into multiple intrinsic mode function signal components;

[0009] Determine the correlation between each intrinsic mode function signal component and the external leakage magnetic signal;

[0010] The top N intrinsic mode function signal components with high correlation are selected for external leakage magnetic signal reconstruction to obtain the reconstructed signal;

[0011] Determine if there is a fault in the wind turbine based on the reconstructed signal.

[0012] In one embodiment, determining whether a wind turbine has a fault based on the reconfiguration signal includes:

[0013] Based on the harmonics of the reconstructed signal, determine the total effective value of the harmonics of the reconstructed signal;

[0014] The presence of a fault in a wind turbine is determined by the change in the total harmonic effective value of the reconstructed signal within a set time period.

[0015] In one embodiment, determining the total harmonic effective value of the reconstructed signal based on its harmonics includes:

[0016] The total harmonic RMS value of the reconstructed signal is determined using the following formula:

[0017]

[0018] Among them, A i Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics.

[0019] In one embodiment, determining whether a wind turbine has a fault based on the change in the total harmonic effective value of the reconstructed signal within a set time period includes:

[0020] The following formula can be used to determine if a wind turbine has a fault:

[0021]

[0022] Where ΔT is the set time period, A t+ΔT A represents the total harmonic RMS value of the reconstructed signal at time t+ΔT. t This represents the total harmonic effective value of the reconstructed signal at time t;

[0023] If c% exceeds the set threshold, the wind turbine is considered to be faulty.

[0024] In one embodiment, the method further includes:

[0025] A genetic algorithm is used to optimize the variational mode decomposition method and determine the optimal variational mode decomposition parameters.

[0026] In one embodiment, determining whether a wind turbine has a fault based on the reconfiguration signal includes:

[0027] The reconstructed signal is subjected to Hilbert-Huang transform to obtain signal features, which include envelope, instantaneous phase, and instantaneous frequency.

[0028] Determine whether a wind turbine has a fault by analyzing the changes in the signal characteristics of the reconstructed signal within a set time period.

[0029] Secondly, a wind turbine generator fault detection device based on magnetic flux leakage detection is provided, comprising:

[0030] External leakage magnetic field signal acquisition module, used to acquire the external leakage magnetic field signal of the wind turbine;

[0031] The signal decomposition module is used to decompose the external leakage magnetic signal into multiple intrinsic mode function signal components using the variational mode decomposition method.

[0032] The correlation determination module is used to determine the correlation between each intrinsic mode function signal component and the external leakage magnetic signal;

[0033] The signal reconstruction module is used to select the top N intrinsic mode function signal components with high correlation to reconstruct the external leakage magnetic signal and obtain the reconstructed signal.

[0034] The fault diagnosis module is used to determine whether there is a fault in the wind turbine based on the reconstructed signal.

[0035] In one embodiment, the fault diagnosis module is further configured to:

[0036] Based on the harmonics of the reconstructed signal, determine the total effective value of the harmonics of the reconstructed signal;

[0037] The presence of a fault in a wind turbine is determined by the change in the total harmonic effective value of the reconstructed signal within a set time period.

[0038] In one embodiment, the fault diagnosis module is further configured to:

[0039] The total harmonic RMS value of the reconstructed signal is determined using the following formula:

[0040]

[0041] Among them, A i Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics.

[0042] In one embodiment, the fault diagnosis module is further configured to:

[0043] The reconstructed signal is subjected to Hilbert-Huang transform to obtain signal features, which include envelope, instantaneous phase, and instantaneous frequency.

[0044] Determine whether a wind turbine has a fault by analyzing the changes in the signal characteristics of the reconstructed signal within a set time period.

[0045] Compared with the prior art, this application has the following advantages: This application monitors the wind turbine based on the external leakage magnetic array signal of the generator and adopts the signal reconstruction method, which can better meet the requirements of non-intrusive, low cost and reliability, and effectively extract and identify the characteristics of short circuit faults between stator and rotor turns of the wind turbine. Attached Figure Description

[0046] This application can be better understood by referring to the description given below in conjunction with the accompanying drawings, which, together with the detailed description below, are incorporated in and form part of this specification. In the drawings:

[0047] Figure 1 A flowchart of a wind turbine fault detection method based on leakage magnetic field detection according to an embodiment of this application is shown;

[0048] Figure 2 A structural block diagram of a wind turbine fault detection device based on leakage magnetic field detection according to an embodiment of this application is shown;

[0049] Figure 3 A schematic diagram showing an external leakage magnetic field sensor arranged on a wind turbine according to an embodiment of this application is shown;

[0050] Figure 4 The variational mode decomposition (VMD) diagram of the 5-turn short-circuit rotor winding is shown.

[0051] Figure 5 The variational mode decomposition (VMD) diagram of the 5-turn short-circuit stator winding is shown.

[0052] Figure 6 The envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal when 5 turns of the stator winding are short-circuited are shown.

[0053] Figure 7 The envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal when 5 turns of the rotor winding are short-circuited are shown.

[0054] Figure 8 The total effective value of leakage magnetic harmonics during rotor inter-turn short-circuit faults is shown;

[0055] Figure 9 The total effective value of the leakage magnetic harmonics during stator inter-turn short-circuit faults is shown. Detailed Implementation

[0056] Exemplary embodiments of the present application will be described below with reference to the accompanying drawings. For clarity and brevity, not all features of the actual embodiments are described in the specification. However, it should be understood that many embodiment-specific decisions can be made in the development of any such actual embodiment to achieve the developer’s specific objectives, and these decisions may vary as the embodiments differ.

[0057] It should also be noted that, in order to avoid obscuring this application with unnecessary details, only the device structure closely related to the solution according to this application is shown in the accompanying drawings, while other details that are not closely related to this application are omitted.

[0058] It should be understood that this application is not limited to the described embodiments by virtue of the following description with reference to the accompanying drawings. In this document, embodiments may be combined with each other, features may be substituted or borrowed between different embodiments, and one or more features may be omitted in one embodiment, where feasible.

[0059] Figure 1 A flowchart of a wind turbine fault detection method based on leakage magnetic field detection according to an embodiment of this application is shown. See also: Figure 1 The methods include:

[0060] Step S110: Obtain the external leakage magnetic signal of the wind turbine. In this step, the wind turbine can be a doubly fed wind turbine. One leakage magnetic sensor is arranged at one end of the wind turbine, and three leakage magnetic sensors are evenly arranged in the radial circumference of the wind turbine to obtain the external leakage magnetic signal detected by each leakage magnetic sensor.

[0061] Step S120: The external leakage magnetic signal is decomposed into multiple intrinsic mode function (IMF) signal components using the variational mode decomposition (VMD) method.

[0062] In this step, a genetic algorithm (GA) can be used to optimize the variational mode decomposition method and achieve automatic parameter selection. Here, the GA parameters are initialized with a maximum number of iterations of 20, a population size of 10, a crossover probability of 0.9, and a mutation probability of 0.1. Through the process of genetic selection, crossover, and mutation, the optimal variational mode decomposition (VMD) parameters are output when the optimal number of iterations is reached.

[0063] Step S130: Determine the correlation between each intrinsic mode function signal component and the external leakage magnetic signal; in this step, the signal component with the strongest correlation can be selected by using the correlation evaluation criterion.

[0064] Step S140: Select the top N intrinsic mode function signal components with high correlation to reconstruct the external leakage magnetic signal and obtain the reconstructed signal; here, N can be 2.

[0065] In this step, the N most relevant intrinsic mode function signal components are added together to reconstruct the external leakage magnetic signal.

[0066] Step S150: Determine whether there is a fault in the wind turbine based on the reconstructed signal.

[0067] The embodiments of this application monitor the wind turbine generator based on the external leakage magnetic array signal of the generator and adopt the signal reconstruction method, which can better meet the requirements of non-intrusive, low cost and reliability, and effectively extract and identify the characteristics of short circuit faults between stator and rotor of the wind turbine generator.

[0068] In one embodiment, determining whether a wind turbine has a fault based on the reconfiguration signal may include:

[0069] Based on the harmonics of the reconstructed signal, the total effective harmonic value (TRM) of the reconstructed signal is determined. In this step, the TRM of the reconstructed signal is determined using the following formula:

[0070]

[0071] Among them, A i Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics. Here, harmonics include integer harmonics and fractional harmonics.

[0072] The presence of a fault in a wind turbine is determined by the change in the total harmonic effective value of the reconstructed signal within a set time period ΔT.

[0073] In this step, the following formula is used to determine whether the wind turbine has a fault:

[0074]

[0075] Among them, A t+ΔT A represents the total harmonic RMS value of the reconstructed signal at time t+ΔT. t This represents the total harmonic effective value of the reconstructed signal at time t;

[0076] If c% exceeds the set threshold, it is determined that there is a fault in the wind turbine; here, the fault of the wind turbine can be, for example, an inter-turn short circuit fault.

[0077] In one embodiment, determining whether a wind turbine has a fault based on the reconfiguration signal may further include:

[0078] The reconstructed signal is subjected to Hilbert-Huang (HHT) transform to obtain signal features, which include envelope, instantaneous phase, and instantaneous frequency.

[0079] The presence of a fault in a wind turbine is determined by analyzing the changes in the signal characteristics of the reconstructed signal over a set time period. In this step, significant changes in the envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal indicate a fault in the wind turbine.

[0080] Based on the same inventive concept as the wind turbine fault detection method based on magnetic flux leakage detection provided in the embodiments of this application, the embodiments of this application also provide a wind turbine fault detection device based on magnetic flux leakage detection. Figure 2 A structural block diagram of a wind turbine fault detection device based on leakage magnetic field detection according to an embodiment of this application is shown. The device includes:

[0081] The external leakage magnetic field signal acquisition module 210 is used to acquire the external leakage magnetic field signal of the wind turbine. The wind turbine can be a doubly fed wind turbine. One external leakage magnetic field sensor is arranged at one end of the wind turbine, and three external leakage magnetic field sensors are evenly arranged in the radial circumference of the wind turbine to acquire the external leakage magnetic field signal detected by each external leakage magnetic field sensor.

[0082] The signal decomposition module 220 is used to decompose the external leakage magnetic signal into multiple intrinsic mode function signal components using the variational mode decomposition method. Here, a genetic algorithm (GA) can be used to optimize the variational mode decomposition method to achieve automatic parameter selection. The GA parameters are initialized with a maximum number of iterations of 20, a population size of 10, a crossover probability of 0.9, and a mutation probability of 0.1. Through the process of genetic selection, crossover, and mutation, the optimal variational mode decomposition VMD parameters are output when the optimal number of iterations is reached.

[0083] The correlation determination module 230 is used to determine the correlation between each intrinsic mode function signal component and the external leakage magnetic signal; here, the correlation evaluation criterion method can be used to select the signal component with the strongest correlation.

[0084] The signal reconstruction module 240 is used to select the top N intrinsic mode function signal components with high correlation to reconstruct the external leakage magnetic signal and obtain the reconstructed signal; here, N can be 2.

[0085] The fault diagnosis module 250 is used to determine whether there is a fault in the wind turbine based on the reconfiguration signal.

[0086] The embodiments of this application monitor the wind turbine generator based on the external leakage magnetic array signal of the generator and adopt the signal reconstruction method, which can better meet the requirements of non-intrusive, low cost and reliability, and effectively extract and identify the characteristics of inter-turn short circuit faults in the stator and rotor of the doubly fed wind turbine generator.

[0087] In one embodiment, the fault diagnosis module 250 is further configured to:

[0088] Based on the harmonics of the reconstructed signal, determine the total effective value of the harmonics of the reconstructed signal;

[0089] The presence of a fault in a wind turbine is determined by the change in the total harmonic effective value of the reconstructed signal within a set time period.

[0090] In one embodiment, the fault diagnosis module 250 is further configured to:

[0091] The total harmonic RMS value of the reconstructed signal is determined using the following formula:

[0092]

[0093] Among them, A i Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics.

[0094] In one embodiment, the fault diagnosis module is further configured to:

[0095] The reconstructed signal is subjected to Hilbert-Huang transform to obtain signal characteristics, including envelope, instantaneous phase, and instantaneous frequency. The changes in the signal characteristics of the reconstructed signal within a set time period are used to determine whether the wind turbine has a fault.

[0096] In this embodiment, when the envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal change significantly, it indicates that there is a fault in the wind turbine.

[0097] To verify the effectiveness of the wind turbine fault detection method and device based on leakage flux detection provided in this application, this application uses an external leakage flux sensor to collect external leakage flux signals of a doubly-fed wind turbine under different stator and rotor inter-turn short-circuit faults on a 100kW three-phase wound rotor doubly-fed induction generator test simulation platform. Figure 3 A schematic diagram of an external magnetic flux leakage sensor arranged on a wind turbine according to an embodiment of this application is shown. The external magnetic flux leakage sensor used in this application adopts an HTANT8001R loop antenna with 36 turns of internal coil and a coil diameter of 13.3 cm. One external magnetic flux leakage sensor is arranged at the end of the doubly fed wind turbine and three are arranged in the radial position. The plane where the external magnetic flux leakage sensor is located is parallel to the end and radial housing of the doubly fed wind turbine, respectively.

[0098] External leakage flux signals of a doubly-fed induction generator (DFIG) were sampled at a sampling rate of 102400 Hz, with 204800 sampling points (preset to 40960), and an inter-turn short-circuit duration of 1 second. Actual tests were conducted on the DFIG under healthy conditions and with different numbers of turns, including stator and rotor inter-turn short-circuit faults. Under grid-connected conditions, the external leakage flux signal sampling frequency of the DFIG was 102400 Hz, the rotational speed was 1800 r / min, and the load was 50 kW. Taking the end leakage flux signal analysis with a 5-turn short circuit in the stator and rotor as an example, Variational Mode Decomposition (VMD) was used to adaptively decompose the external leakage flux signal into 7 intrinsic mode function signal components. Figure 4 The variational mode decomposition (VMD) plot of the 5-turn short-circuit rotor winding is shown. Figure 5 The variational mode decomposition (VMD) diagram of the 5-turn short-circuit stator winding is shown. Then, based on the evaluation criteria of the correlation coefficient, the best signal component with the most obvious characteristics is selected from the signal components of the intrinsic mode function, and the signal component with the highest correlation is selected to reconstruct the external leakage magnetic signal.

[0099] Signal components sensitive to external leakage flux characteristics were selected for signal reconstruction. Hilbert-Huang transform was used to analyze the stator and rotor short-circuit reconstructed signals. Figure 6 The envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal when 5 turns of the stator winding are short-circuited are shown. Figure 7 The envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal when 5 turns of the rotor winding are short-circuited are shown, including the envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal under both healthy and fault conditions. The envelope, instantaneous phase, and instantaneous frequency of the reconstructed signal all change significantly when a fault occurs. This characteristic phenomenon is quite significant and can be used as a fault characteristic signal for fault diagnosis and identification of stator and rotor inter-turn short circuits. Table 1 shows the results under healthy and fault conditions.

[0100] Table 1 Results of Fault and Health Conditions

[0101]

[0102] Further analysis of external leakage magnetic signals at different locations of the doubly-fed wind turbine generator Figure 8 The total effective value of the leakage magnetic harmonics during rotor inter-turn short-circuit faults is shown. Figure 9The table shows the total effective value of the leakage magnetic harmonics in a stator inter-turn short-circuit fault. At different axial and radial positions, the total effective value A increases with the severity of the inter-turn short-circuit fault. Comparison of sensors at different positions shows that radial positions 1 and 3 are symmetrical, and the monitoring of leakage magnetic signals is essentially the same. Although the total effective value A of the leakage magnetic signal at the axial position of the generator end is slightly lower than that at radial positions 1 and 3, the response of the leakage magnetic signal becomes more pronounced with the increase of the fault severity. The characteristic frequencies of the stator and rotor principal components show good consistency with the total effective value A, indicating that at different axial and radial positions, the total effective value of each harmonic component of the leakage magnetic signal increases with the increase of the short-circuit fault severity, which can be used to diagnose short-circuit faults in wind turbine windings. Table 2 shows the total effective value A of the rotor inter-turn short-circuit fault at different axial and radial positions with different fault severity, and Table 3 shows the total effective value A of the stator inter-turn short-circuit fault at different axial and radial positions with different fault severity.

[0103] Table 2 Comparison of stator inter-turn short circuit faults with different locations and number of short turns

[0104]

[0105] Table 3 Comparison of faults with different short-turn numbers at different locations of rotor inter-turn short circuits

[0106]

[0107]

[0108] The above descriptions are merely various embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting a fault of a wind power generator based on magnetic flux leakage detection, characterized in that, include: Acquire the external leakage magnetic field signal of the wind turbine; The external leakage magnetic signal is decomposed into multiple intrinsic mode function signal components using the variational mode decomposition method. Determine the correlation between each of the intrinsic mode function signal components and the external leakage magnetic signal; The top N intrinsic mode function signal components with high correlation are selected for external leakage magnetic signal reconstruction to obtain the reconstructed signal; Determining whether the wind turbine has a fault based on the reconstructed signal includes: Based on the harmonics of the reconstructed signal, determine the total harmonic effective value of the reconstructed signal; The wind turbine is determined to have a fault based on the change in the total harmonic effective value of the reconstructed signal within a set time period. Based on the harmonics of the reconstructed signal, the total harmonic effective value of the reconstructed signal is determined, including: The total harmonic RMS value of the reconstructed signal is determined using the following formula: A= in, Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics. Determining whether the wind turbine has a fault based on the change in the total harmonic effective value of the reconstructed signal within a set time period includes: The following formula can be used to determine if a wind turbine has a fault: in, To set a time period, Indicates the first The total harmonic RMS value of the reconstructed signal at time t. Indicates the first The effective value of the total harmonics of the reconstructed signal at time t; like If the threshold is exceeded, the wind turbine is considered to be faulty.

2. The method as described in claim 1, characterized in that, The method further includes: A genetic algorithm is used to optimize the variational mode decomposition method and determine the optimal variational mode decomposition parameters.

3. The method as described in claim 1, characterized in that, in, Determining whether the wind turbine has a fault based on the reconstructed signal includes: The reconstructed signal is subjected to Hilbert-Huang transform to obtain signal features, which include envelope, instantaneous phase, and instantaneous frequency. The wind turbine is determined to have a fault based on the changes in the signal characteristics of the reconstructed signal within a set time period.

4. A wind turbine generator fault detection device based on magnetic flux leakage detection, characterized in that, include: External leakage magnetic field signal acquisition module, used to acquire the external leakage magnetic field signal of the wind turbine; The signal decomposition module is used to decompose the external leakage magnetic signal into multiple intrinsic mode function signal components using the variational mode decomposition method; A correlation determination module is used to determine the correlation between each of the intrinsic mode function signal components and the external leakage magnetic signal; The signal reconstruction module is used to select the top N intrinsic mode function signal components with high correlation to reconstruct the external leakage magnetic signal and obtain the reconstructed signal. The fault diagnosis module is used to determine whether the wind turbine has a fault based on the reconstructed signal; The fault diagnosis module is also used for: Based on the harmonics of the reconstructed signal, determine the total harmonic effective value of the reconstructed signal; Determining whether the wind turbine has a fault based on the change in the total harmonic effective value of the reconstructed signal within a set time period includes: The following formula can be used to determine if a wind turbine has a fault: in, To set a time period, Indicates the first The total harmonic RMS value of the reconstructed signal at time t. Indicates the first The effective value of the total harmonics of the reconstructed signal at time t; like If the threshold is exceeded, the wind turbine is considered to be faulty. The fault diagnosis module is also used for: The total harmonic RMS value of the reconstructed signal is determined using the following formula: A= in, Let M be the i-th harmonic of the reconstructed signal, and M be the number of harmonics.

5. The apparatus as described in claim 4, characterized in that, in, The fault diagnosis module is also used for: The reconstructed signal is subjected to Hilbert-Huang transform to obtain signal features, which include envelope, instantaneous phase, and instantaneous frequency. The wind turbine is determined to have a fault based on the changes in the signal characteristics of the reconstructed signal within a set time period.

Images