Abnormal detection device for transmission chain of wind turbine generator

Abstract

The invention belongs to the technical field of wind turbine generator detection, particularly relates to a wind turbine generator transmission chain anomaly detection device, and provides the following scheme for solving the problems that an existing installation mode is loose and missed in detection, magnetorheological fluid single functional application is achieved, and multi-action cooperative control is lacked. Comprising a positioning base, a detection base, a current sensor and an assembly sensor assembly, an electromagnet is embedded in the bottom surface of the positioning base, an overall framework of the wind turbine generator transmission chain abnormity detection device is constructed, magnetic adsorption fixation is provided through the electromagnet embedded in the positioning base, and under the action of a magnetic field of the electromagnet, the current sensor is arranged on the current sensor. Rigid locking is formed between the sliding block and the sliding groove and between the butt joint rod and the butt joint groove, the mechanical connection reliability is ensured, and it is ensured that the device is stably installed on the metal surface of the wind turbine generator transmission chain; and a plurality of sensors are arranged to carry out multi-physical quantity cooperative monitoring, and a transmission chain holographic monitoring network is constructed through space integration and time synchronization of vibration, temperature, noise and current sensors.

Description

Technical Field

[0001] This invention relates to a detection device, specifically a wind turbine drive chain anomaly detection device, belonging to the field of wind turbine detection technology. Background Technology

[0002] As the core hub of energy conversion, the health of the wind turbine drivetrain directly determines the unit's power generation efficiency and operational safety. Key components in the drivetrain, such as the main shaft, gearbox bearings, and high-speed shaft, are subjected to complex alternating loads and extreme environmental stresses over long periods, making them prone to progressive failures such as bearing wear, gear pitting, and shaft misalignment. Traditional maintenance methods rely on periodic inspections and reactive repairs, which are not only slow to respond but also fail to detect early, subtle fault characteristics, leading to worsening of the faults and even cascading shutdowns.

[0003] Existing technologies generally employ a combination of triaxial vibration sensors, temperature sensors, and current sensors, attempting to diagnose faults by correlating multiple parameters such as vibration spectrum, temperature rise curve, and current harmonics. For example, CN220378407U discloses a wind turbine transmission chain fault detection device based on displacement, bending moment, and torque. Displacement sensors are installed on the wind turbine main shaft, gearbox high-speed shaft, generator drive end, and generator non-drive end; bending moment and torque sensors are installed on the wind turbine main shaft and connected via a data acquisition unit. However, the sensor installation method severely restricts data reliability. Each sensor is installed independently, lacking a spatiotemporal calibration mechanism. For instance, localized overheating of the gearbox and time deviation of the meshing vibration signal may be misjudged as asynchronous faults, reducing diagnostic accuracy. A single vibration signal cannot comprehensively reflect the fault state. For example, gearbox faults are accompanied by sudden changes in vibration energy and abnormal temperatures. Furthermore, CN217883119U discloses a wind power generation system with a magnetorheological fluid transducer. The wind power generation system includes an impeller, a magnetorheological fluid transducer, and a synchronous generator. The impeller is rotated by wind energy. A magnetorheological fluid (MRF) generator includes a driving impeller, a driven impeller, and a coil. The driving and driven impellers are coaxial and opposite to each other, with a gap between them filled with MRF. An impeller is driven by the driving impeller to rotate, and the driving impeller drives the driven impeller to rotate via the MRF. The coil is fitted into the gap, and when energized, it generates a magnetic field. The MRF changes its density and viscosity under the influence of the magnetic field, thus changing the speed ratio of the MRF generator and keeping the driven impeller's rotational speed constant. While existing technologies also involve utilizing the properties of MRF in wind power systems, they only involve adjusting... While focusing on power transmission by altering the viscosity of a liquid using a magnetic field to achieve gear ratio control, this approach overlooks its innovative value in structurally fixed scenarios. Traditional vibration sensors require rigid fixation to the equipment surface, but the surfaces of transmission chain components (such as spindle bearings and high-speed shafts of gearboxes) are often curved or confined spaces, resulting in insufficient sensor installation stability. Vibration sensors need to be rigidly fixed to the equipment surface by welding or bolts, but the bolt preload is easily attenuated by vibration and impact, causing micron-level gaps between the sensor and the base. For example, bolt-fastened bases are prone to loosening under vibration and impact, causing signal drift or distortion. Summary of the Invention

[0004] This invention provides a wind turbine drivetrain anomaly detection device to address the problems of loose installation and missed detection in traditional detection devices, the single-function application of magnetorheological fluid, and the lack of multi-action coordinated control.

[0005] The present invention achieves the above objectives through the following technical solution: a wind turbine drive chain abnormality detection device, comprising a wind turbine drive chain and a detection component, the detection component comprising a positioning base, a detection base, a current sensor and an assembly sensor assembly, the detection base and the current sensor being movably connected to the positioning base, the detection base being fixedly connected to the wind turbine drive chain, the assembly sensor assembly being movably connected to the detection base, the assembly sensor assembly including but not limited to a triaxial vibration sensor, a temperature sensor and a noise sensor, and an electromagnet embedded in the bottom surface of the positioning base; The detection base includes multiple sensor mounting bases, with a docking rod movably connected between two adjacent sensor mounting bases. A deformation bladder is connected to the bottom surface of the sensor mounting base, and the deformation bladder is filled with magnetorheological fluid. A conical fixing hole is opened in the middle part of the sensor mounting base, and a conical fixing sleeve is movably installed in the conical fixing hole. The positioning base has sliding grooves on both sides, and multiple sliders are movably connected on the positioning base. The sensor fixing base has docking grooves at both ends. Multiple self-locking pins are embedded in the bottom surface of the sliding grooves and the inner wall of the docking grooves. The self-locking pins are filled with magnetorheological fluid.

[0006] As a further embodiment of the present invention: the wind turbine transmission chain includes a turbine housing, a generator is fixedly connected inside the turbine housing, a bushing is provided on the generator's motor shaft sleeve, a gearbox shaft sleeve is provided with a bearing and coaxially connected to the generator's motor shaft, a detection base is fixedly connected to the bushing and the bearing respectively, a yaw motor is also connected inside the turbine housing, and current sensors pass through the lines of the generator and the yaw motor respectively.

[0007] As a further embodiment of the present invention: the bottom surface of the positioning base is also embedded with a permanent magnet, and the two sides of the slider are fixedly connected with limit blocks. The limit blocks are movably locked in the slide groove. The bottom surface of the limit blocks is provided with positioning slots that are equidistantly arranged. Multiple self-locking pins embedded in the bottom surface of the slide groove are equidistantly distributed, and the distance between two adjacent positioning slots is equal to the distance between adjacent self-locking pins embedded in the bottom surface of the slide groove. The bottom surfaces of the limit blocks are provided with arc-shaped end faces at both ends.

[0008] As a further embodiment of the present invention: a self-locking support rod is connected between the probe base and the slider. The self-locking support rod includes an outer tube and a telescopic inner rod. The bottom end of the telescopic inner rod is movably inserted into the outer tube, and multiple retaining rings are fixedly connected to one end of the telescopic inner rod located in the outer tube. A support spring is provided in the outer tube. In the initial state, the support spring is compressed and presses against the bottom of the telescopic inner rod. The outer tube is also filled with magnetorheological fluid.

[0009] As a further embodiment of the present invention: an inner groove is provided in the middle part of the bottom surface of the sensor fixing base, and the conical fixing sleeve is inserted into the conical fixing hole from the inner groove. The sleeve body of the conical fixing sleeve has cross-shaped grooves.

[0010] As a further embodiment of the present invention: two docking rods are connected between two adjacent sensor mounting bases, one of which is connected to a ball socket and the other is connected to a universal ball, and the universal ball is movably embedded in the ball socket.

[0011] As a further embodiment of the present invention: the inner wall of the docking groove is symmetrically provided with multiple sets of self-locking pins, and an annular docking slot is provided on one end of the docking rod inserted into the docking groove. The annular docking slot is engaged with one of the sets of self-locking pins, and a tapered end is provided on one end of the docking rod inserted into the docking groove.

[0012] As a further embodiment of the present invention: the self-locking pin includes a pin housing and a pin inner rod. The rod body of the pin inner rod is movably inserted into the pin housing. An annular groove is formed on the rod body of the pin inner rod located inside the pin housing. A sealing ring is fitted inside the annular groove. An arc-shaped end is provided on the outer end of the pin inner rod. A locking spring is provided inside the pin housing, and the locking spring is compressed in the initial state and rests against the bottom of the pin inner rod.

[0013] As a further aspect of the present invention: a buffer cavity is provided on the side wall of the pin housing, and a plurality of guide holes are connected between the inner cavity of the pin housing and the buffer cavity, and the guide holes are located at the bottom end of the side wall of the pin housing.

[0014] As a further aspect of the present invention: the current sensor and each sensor in the assembled sensor assembly are connected to a peripheral data acquisition device. The data acquisition device collects the high-frequency data from the current sensor and each sensor in the assembled sensor assembly and performs preliminary filtering. Then, the data is sent to an edge computing terminal. The edge terminal has a built-in high-speed computing chip and storage unit, and runs spectrum analysis and machine learning algorithms in real time to perform real-time calculation and storage processing on the data. The spectrum analysis and machine learning algorithms include: Offline benchmark model training: Using selected equipment health status data, a benchmark model is trained that includes the target quantity estimate and its reference fluctuation range calculation. Online benchmark model prediction and health assessment utilizes online data of various factors to continuously calculate the estimated values ​​of target quantities and compare them with the measured values ​​to assess the deviation between the two. Furthermore, based on the deviation between the measured values ​​and the dynamic range given by the benchmark model, the health status of the wind turbine is assessed. When the deviation is significant to a certain extent, it is considered that the wind turbine may be in an abnormal state, thereby issuing an early warning.

[0015] The beneficial effects of this invention are: 1. This invention comprises a positioning base, a detection base, a current sensor, and an assembly sensor component. The assembly sensor component includes, but is not limited to, a triaxial vibration sensor, a temperature sensor, and a noise sensor. An electromagnet is embedded in the bottom surface of the positioning base. The positioning base, detection base, current sensor, and assembly sensor component together form the overall architecture of the wind turbine drivetrain anomaly detection device. The electromagnet embedded in the positioning base provides magnetic adsorption fixation, ensuring stable installation of the device on the metal surface of the wind turbine drivetrain. Furthermore, multiple sensors are used for collaborative monitoring of multiple physical quantities. Through spatial integration and temporal synchronization of vibration, temperature, noise, and current sensors, a holographic monitoring network for the drivetrain is constructed. This network can monitor the health status of the drivetrain in real time and predict fault risks, which is of great significance for improving the stability and economy of wind turbines. 2. The detection base of this invention includes multiple sensor mounting bases, docking rods, and deformation bladders. The deformation bladders are filled with magnetorheological fluid. A conical fixing hole is provided in the middle of the sensor mounting base, and a conical fixing sleeve is movably installed in the conical fixing hole. The detection base adopts a modular design, which can flexibly adjust the sensor layout according to the transmission chain structure, and can assemble a corresponding number of sensor mounting bases for different detection parts of the transmission chain. The deformation bladder is filled with magnetorheological fluid, which can adapt to the curvature of the component surface and solidify and lock it. The conical fixing hole and the conical fixing sleeve cooperate to realize the quick assembly and disassembly of the sensor assembly. When the electromagnet is energized, it can fix the positioning base. At the same time, the magnetic field generated by the energized electromagnet can also turn the magnetorheological fluid filled in the deformation bladder into a solid-like state, thereby ensuring the further locking of the detection base to the detection part. 3. This invention features multiple self-locking pins embedded in the bottom surface of the chute and the inner wall of the docking groove. These self-locking pins are filled with magnetorheological fluid. Utilizing the properties of the magnetorheological fluid, the self-locking pins facilitate the movement of the slider on the chute, enabling the connection between the detection base and the detection part, and facilitating the assembly of adjacent sensor mounting bases. Furthermore, under the magnetic field generated by the electromagnet, a rigid lock is formed between the slider and the chute, and between the docking rod and the docking groove, ensuring reliable mechanical connections. This eliminates the need for welding or bolts to secure the connections, significantly improving construction efficiency, solving the problems of poor adaptability and easy loosening in traditional sensor installations, and enhancing the accuracy of monitoring data and system stability. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall installation structure of the detection device of the present invention; Figure 2 This is a schematic diagram of the composition and structure of the detection device of the present invention; Figure 3 This is a schematic diagram of the connection structure between the positioning base and the slider of the present invention; Figure 4This is a schematic diagram of the cross-sectional structure of the positioning base of the present invention; Figure 5 For the present invention Figure 4 Schematic diagram of the structure at point A in the middle; Figure 6 This is a schematic diagram of the cross-sectional structure of the slider of the present invention; Figure 7 This is a schematic diagram of the cross-sectional structure of the self-locking strut of the present invention; Figure 8 This is a schematic diagram of the detection base structure of the present invention; Figure 9 This is a schematic diagram of the connection structure between the sensor mounting base and the docking rod of the present invention; Figure 10 This is a schematic cross-sectional view of the sensor mounting base of the present invention; Figure 11 This is a schematic diagram of the combined connection structure of the two connecting rods of the present invention; Figure 12 This is a schematic diagram of the conical fixing sleeve structure of the present invention; Figure 13 This is a schematic diagram of the cross-sectional structure of the self-locking pin of the present invention; Figure 14 This is a schematic diagram of the frequency domain analysis principle of embodiment three of the present invention; Figure 15 Vibration spectrum of Embodiment 3 of the present invention Figure 1 ; Figure 16 Vibration spectrum of Embodiment 3 of the present invention Figure 2 ; Figure 17 Vibration spectrum of Embodiment 3 of the present invention Figure 3 ; Figure 18 This is a schematic diagram of a data sample from Embodiment 3 of the present invention.

[0017] In the diagram: 1. Generator casing; 11. Generator; 12. Bushing; 13. Bearing; 2. Positioning base; 21. Slide groove; 22. Permanent magnet; 23. Electromagnet; 3. Slider; 31. Limiting block; 32. Positioning slot; 33. Arc-shaped end face; 4. Self-locking support rod; 41. Outer tube; 42. Telescopic inner rod; 43. Retaining ring; 44. Support spring; 5. Detector base; 51. Sensor mounting base; 52. Connecting rod; 53. Inner groove; 54. Conical fixing sleeve 55. Deformation bladder; 56. Conical fixing hole; 57. Docking groove; 58. Annular docking slot; 59. Conical end; 510. Ball socket seat; 511. Universal ball; 512. Gap groove; 6. Self-locking pin; 61. Pin housing; 62. Pin inner rod; 63. Arc end; 64. Annular groove; 65. Sealing ring; 66. Limiting inner arc surface; 67. Locking spring; 68. Buffer cavity; 69. Guide hole; 7. Current sensor; 8. Assemble sensor assembly. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 like Figures 1 to 13 As shown, a wind turbine drivetrain anomaly detection device includes a wind turbine drivetrain and a detection component. The detection component includes a positioning base 2, a detection base 5, a current sensor 7, and an assembly sensor assembly 8. The detection base 5 and the current sensor 7 are movably connected to the positioning base 2, and the detection base 5 is fixedly connected to the wind turbine drivetrain. The assembly sensor assembly 8 is movably connected to the detection base 5. The assembly sensor assembly 8 includes, but is not limited to, a triaxial vibration sensor, a temperature sensor, and a noise sensor. An electromagnet 23 is embedded in the bottom surface of the positioning base 2. The overall architecture of the wind turbine drivetrain anomaly detection device is constructed through the positioning base 2, the detection base 5, the current sensor 7, and the assembly sensor assembly 8. The electromagnet 23 embedded in the positioning base 2 provides magnetic adsorption fixation, ensuring stable installation of the device on the metal surface of the wind turbine drivetrain. Furthermore, multiple sensors are set up for multi-physical quantity collaborative monitoring. Through spatial integration and temporal synchronization of vibration, temperature, noise, and current sensors, a holographic monitoring network for the drivetrain is constructed, which can monitor the health status of the drivetrain in real time and predict fault risks, which is of great significance for improving the stability and economy of wind turbines. The detection base 5 includes multiple sensor mounting bases 51. A docking rod 52 is movably connected between two adjacent sensor mounting bases 51. A deformation bladder 55 is connected to the bottom surface of the sensor mounting base 51. The deformation bladder 55 is filled with magnetorheological fluid. A conical fixing hole 56 is opened in the middle part of the sensor mounting base 51. A conical fixing sleeve 54 is movably installed in the conical fixing hole 56. The detection base 5 adopts a modular design, which can flexibly adjust the sensor layout according to the transmission chain structure. It can also assemble a corresponding number of sensor mounting bases 51 for different detection parts of the transmission chain. The deformation bladder 55 is filled with magnetorheological fluid, which can adapt to the curvature of the component surface and solidify and lock it. The conical fixing hole 56 and the conical fixing sleeve 54 cooperate to realize the quick assembly and disassembly of the sensor assembly 8. When the electromagnet 23 is energized, it can fix the positioning base 2. At the same time, the magnetic field generated by the electromagnet 23 can also turn the magnetorheological fluid filled in the deformation bladder 55 into a solid-like state, thereby ensuring the further locking of the detection base 5 to the detection part. The positioning base 2 has sliding grooves 21 on both sides, and multiple sliders 3 are movably connected to the positioning base 2. The sensor fixing seat 51 has docking grooves 57 at both ends. Multiple self-locking pins 6 are embedded in the bottom surface of the sliding groove 21 and the inner wall of the docking groove 57. The self-locking pins 6 are filled with magnetorheological fluid. The self-locking pins 6 utilize the properties of magnetorheological fluid to facilitate the movement of sliders 3 on the sliding groove 21, so as to connect the detection base 5 with the detection part and facilitate the assembly of adjacent sensor fixing seats 51. Under the action of the magnetic field generated by the electromagnet 23, the sliders 3 and the sliding groove 21, and the docking rod 52 and the docking groove 57 can form a rigid lock, ensuring the reliability of the mechanical connection. The fastening of each connection part can be achieved without the need for welding or bolts, which greatly improves the construction efficiency, solves the problems of poor adaptability and easy loosening of traditional sensor installation, and improves the accuracy of monitoring data and system stability.

[0020] Example 2 Improvements based on Example 1: like Figures 1 to 6 As shown, the wind turbine transmission chain includes a turbine housing 1, a generator 11 fixedly connected inside the turbine housing 1, a bushing 12 on the motor shaft of the generator 11, and a bearing 13 on the gearbox shaft coaxially connected to the motor shaft of the generator 11. Detection bases 5 are fixedly connected to the bushing 12 and the bearing 13 respectively. A yaw motor is also connected inside the turbine housing 1. Current sensors 7 are connected to the generator 11 and the yaw motor respectively, ensuring that each sensor in the sensor assembly 8 directly captures the mechanical status of key components such as the generator 11 and the gearbox shaft. The current sensors 7 simultaneously monitor the current of the generator 11 and the yaw motor, correlate mechanical load and electrical parameters, and achieve full coverage monitoring of core components such as the transmission chain main shaft, gearbox, and generator, providing accurate location basis for fault tracing.

[0021] Furthermore, a permanent magnet 22 is embedded in the bottom surface of the positioning base 2. Limiting blocks 31 are fixedly connected to both sides of the slider 3. The limiting blocks 31 are movably locked in the slide groove 21. The bottom surface of the limiting blocks 31 has positioning slots 32 that are equidistantly arranged. Multiple self-locking pins 6 embedded in the bottom surface of the slide groove 21 are equidistantly distributed, and the distance between two adjacent positioning slots 32 is equal to the distance between adjacent self-locking pins 6 embedded in the bottom surface of the slide groove 21. The bottom surface of the limiting blocks 31 has arc-shaped end faces 33 at both ends. Through the complementary advantages of the permanent magnet 22 and the electromagnet 23, The positioning base 2 has an adsorption capacity, which means that the positioning base 2 can be initially fixed by the permanent magnet 22, and then the connection is further strengthened by energizing the electromagnet 23 to enhance the vibration resistance; and the slider 3 can be precisely locked in multiple positions by the cooperation of the limiting block 31 and the equidistant self-locking pin 6 of the slide groove 21. The arc end face 33 reduces the movement resistance of the slider 3, and the self-locking pin 6 can be locked in the positioning slot 32 to lock the slider 3, which greatly improves the positioning reliability of the slider 3 in the vibration environment and simplifies the installation and adjustment process.

[0022] like Figure 1 , Figure 2 and Figure 7 As shown, a self-locking support rod 4 connects the detection base 5 and the slider 3. The self-locking support rod 4 includes an outer tube 41 and a telescopic inner rod 42. The bottom end of the telescopic inner rod 42 is movably inserted into the outer tube 41, and multiple retaining rings 43 are fixedly connected to one end of the telescopic inner rod 42 located inside the outer tube 41. A support spring 44 is provided inside the outer tube 41. In the initial state, the support spring 44 is compressed and presses against the lower part of the telescopic inner rod 42. The outer tube 41 is also filled with magnetorheological fluid. The self-locking support rod 4 connects the detection base 5 and the slider 3. The retaining rings 43 of the telescopic inner rod 42 and the magnetorheological fluid in the outer tube 41 form a damping mechanism. The support spring 44 provides an initial preload, forming initial support for the detection base 5. The magnetic field generated by the electromagnet 23 can also convert the magnetorheological fluid filled in the outer tube 41 into a solid-like state, so that the self-locking support rod 4 can further stabilize the support of the detection base 5.

[0023] like Figure 1 , Figure 2 , Figures 8 to 12 As shown, the bottom surface of the sensor mounting base 51 has an inner groove 53 in the middle. The conical mounting sleeve 54 is inserted into the conical mounting hole 56 through the inner groove 53. The body of the conical mounting sleeve 54 has cross-shaped slots 512. After the sensor is inserted into the conical mounting sleeve 54, the conical mounting sleeve 54 is then inserted into the conical mounting hole 56 through the inner groove 53. The cross slots 512 compress the sensor radially, tightly holding it in place, and also making the conical mounting sleeve 54 more firmly connected in the conical mounting hole 56, eliminating the assembly gap between the sensor and the base.

[0024] Furthermore, two docking rods 52 are connected between two adjacent sensor mounting bases 51. One docking rod 52 is connected to a ball socket 510, and the other docking rod 52 is connected to a universal ball 511. The universal ball 511 is movably embedded in the ball socket 510. The docking rods 52 are hinged to the ball socket 510 by the universal ball 511, which allows the adjacent sensor mounting bases 51 to deflect with multiple degrees of freedom, adapting to non-linear arrangement of the transmission chain, such as the curved surface of the bearing, and ensuring the deformation freedom of the modular detection base 5.

[0025] Furthermore, the inner wall of the docking groove 57 is symmetrically provided with multiple sets of self-locking pins 6. The docking rod 52 is inserted into the docking groove 57 at one end of the rod with an annular docking groove 58, and the annular docking groove 58 is engaged with one of the sets of self-locking pins 6. The end of the docking rod 52 inserted into the docking groove 57 is provided with a conical end 59. The docking rod 52 can be easily guided into the docking groove 57 through the conical end 59. The annular docking groove 58 and the self-locking pins 6 are engaged to achieve a rigid connection. According to the diameter of the installation part, the annular docking groove 58 can be engaged with any matching set of self-locking pins 6, which can further fine-tune the deformation of the detection base 5, simplify the module assembly process, and ensure the tensile and shear strength of the docking part.

[0026] like Figure 1 , Figure 5 , Figure 10 and Figure 13 As shown, the self-locking pin 6 includes a pin housing 61 and a pin inner rod 62. The rod body of the pin inner rod 62 is movably inserted into the pin housing 61. An annular groove 64 is formed on the rod body of the pin inner rod 62 located inside the pin housing 61. A sealing ring 65 is fitted inside the annular groove 64. An arc-shaped end 63 is provided at the outer end of the pin inner rod 62. A locking spring 67 is provided inside the pin housing 61. In the initial state, the locking spring 67 is compressed and pressed against the bottom of the pin inner rod 62. The pin inner rod 62 of the self-locking pin 6 is pushed by the locking spring 67, which allows the arc-shaped end 63 to lock the connected part, achieving initial fixation. The magnetorheological fluid is filled and solidified under the magnetic field to ensure locking, achieving dual-mode locking and fixation.

[0027] Furthermore, a buffer cavity 68 is provided on the side wall of the pin housing 61. Several guide holes 69 are connected between the inner cavity of the pin housing 61 and the buffer cavity 68. The guide holes 69 are located at the bottom end of the side wall of the pin housing 61. The buffer cavity 68 is connected to the inner cavity of the pin housing 61 through the guide holes 69. When the magnetorheological fluid is pressurized, it can overflow and buffer, avoiding the sudden increase of hydraulic pressure during the locking process from damaging the sealing ring 65 and extending the service life of the mechanism.

[0028] Example 3 like Figures 14 to 18As shown, a wind turbine transmission chain anomaly detection device is used on 33 wind turbines in a wind farm in Shanxi Province. Each turbine has triaxial vibration sensors installed on key components such as the main shaft bearing, high and low speed shafts of the gearbox, high speed shaft of the gearbox, and generator bearings. Temperature and noise sensors are also deployed to monitor the equipment's operating status in real time from multiple dimensions. High-frequency data collected by the sensors is aggregated by the acquisition module and transmitted to an edge computing terminal for spectrum analysis and vibration feature extraction. The analysis results and raw data from the edge terminal are transmitted to the wind farm's central control room via the wind turbine ring network. The host computer monitoring system deployed in the central control room receives, stores, and analyzes the data in real time. Current sensor 7 and each sensor in the sensor assembly 8 are connected to an external data acquisition device. The data acquisition device collects the high-frequency data from current sensor 7 and each sensor in the sensor assembly 8, performs preliminary filtering, and then sends the data to the edge computing terminal. The edge terminal has a built-in high-speed computing chip and storage unit, and runs spectrum analysis and machine learning algorithms in real time to perform real-time calculation and storage processing on the data. The spectrum analysis and machine learning algorithms include: Offline benchmark model training: Using selected equipment health status data, a benchmark model is trained that includes the target quantity estimate and its reference fluctuation range calculation. Online benchmark model prediction and health assessment utilizes online data of various factors to continuously calculate the estimated values ​​of target quantities and compare them with the measured values ​​to assess the deviation between the two. Furthermore, based on the deviation between the measured values ​​and the dynamic range given by the benchmark model, the health status of the wind turbine is assessed. When the deviation is significant to a certain extent, it is considered that the wind turbine may be in an abnormal state, thereby issuing an early warning.

[0029] Furthermore, spectral analysis and machine learning algorithms calculate the vibration time-domain and frequency-domain characteristics, specifically including: Vibration time-domain characteristics mainly refer to the statistical characteristics of time-domain signal waveforms. When equipment enters an abnormal or faulty state, its vibration mode often changes, which can be reflected in the changes in signal statistical characteristics. The vibration signal time-domain characteristics used in this invention mainly include mean, standard deviation, peak-to-peak value, and root mean square, as detailed below: (1) Mean The formula for calculating the signal mean is:

[0030] The mean is the statistical average of the first moment of a signal, representing the central tendency of a random process. Random processes cluster and fluctuate around it, and it is the static component of the random process. Here, it reflects the equilibrium point of mechanical vibration. The advantage of using the mean for fault diagnosis is that the detected value is more stable than the peak value.

[0031] (2) Standard deviation The formula for calculating the standard deviation of a signal is:

[0032] Standard deviation is a measure of how dispersed the mean of a set of data is. A large standard deviation indicates that most values ​​differ significantly from the mean; a small standard deviation indicates that the values ​​are closer to the mean.

[0033] (3) Peak-to-peak value The formula for calculating the peak-to-peak value of the signal is:

[0034] Peak-to-peak value refers to the difference between the highest and lowest values ​​of a signal within one period, reflecting the degree to which the signal deviates from its average value. In this context, it describes the range of vibration amplitude variation within a measurement time period.

[0035] (4) Root mean square The formula for calculating the root mean square value of a signal is:

[0036] The root mean square (RMS) value is used to describe the energy of vibration and is the second-order average moment of the signal. Also known as the effective value, the RMS value indicates the signal's power transmission capability. It has a good correlation with abnormal irregular vibration waveforms caused by defects such as bearing surface ripples and is an important indicator used by mechanical fault diagnosis systems to determine whether the operating condition is normal.

[0037] The vibration frequency domain characteristics are calculated using Fourier transform to determine the vibration amplitude spectrum.

[0038] Fourier transform is the most basic spectrum analysis method. Its basic principle is shown in the figure. By converting the time domain signal into a frequency domain distribution, the amplitude-frequency distribution diagram in the frequency domain is obtained, and the fundamental frequency and various harmonics of the device are obtained. Each type of spectrum represents a certain operating state of the device.

[0039] With L 2 (0, 2π) denotes the space of square-integrable functions with a period of 2π. Then for any f ∈ L 2 The Fourier transform (0, 2π) has three fundamental formulas: (1) Fourier series expression:

[0040] (2) Fourier coefficients:

[0041] (3) The two are connected by the Parseval identity:

[0042] Through a pilot application on 33 wind turbines at a wind farm in Shanxi Province, the wind turbine drivetrain anomaly detection device demonstrated outstanding performance and application value. Integrating multi-axis vibration sensors, temperature sensors, and other hardware, along with spectrum analysis and machine learning algorithms, the device achieves efficient data transmission and real-time analysis through edge computing and the wind turbine ring network. Pilot results show that the device can accurately capture changes in the operating status of key components, identify early anomalies, provide accurate health assessments and fault predictions, and effectively support preventative maintenance and optimized operation and maintenance decisions.

[0043] The successful application of this device has significantly improved the operational reliability and maintenance efficiency of wind turbines, reduced the risk of unplanned downtime and maintenance costs, and enabled a shift from traditional passive maintenance to proactive preventative maintenance. Simultaneously, the research findings provide crucial technical support for the digital transformation and intelligent management of wind turbines, positively promoting the efficient operation and safe development of the wind power industry. In the future, the system can be further optimized in its algorithm model and expanded to larger-scale wind farm applications, making a greater contribution to the intelligent operation and maintenance of wind turbines and the development of green energy.

[0044] Working principle: The positioning base 2 is magnetically fixed by the electromagnet 23 embedded in it, and the positioning base 2 is initially fixed by the permanent magnet 22. The slider 3 moves on the slide groove 21 to connect the detection base 5 to the detection part. After the sensor is inserted into the conical fixing sleeve 54, the conical fixing sleeve 54 is embedded into the conical fixing hole 56 through the inner groove 53. The cross slot 512 compresses and tightens the sensor radially, making the conical fixing sleeve 54 more firmly connected in the conical fixing hole 56. The docking rod 52 can be easily guided into the docking groove 57 through the conical end 59. The annular docking slot 58 engages with the self-locking pin 6 to achieve a rigid connection. According to the diameter of the installation part, the annular docking slot 58 can be engaged with any set of self-locking pins 6 to further fine-tune the deformation of the detection base 5, ensuring that the detection base 5 is tightly connected to the detection part. Then, by adjusting the electromagnet 23... The energization further enhances the robustness of the connection points, ensuring stable installation of the device on the metal surface of the wind turbine drive chain. Simultaneously, the magnetic field generated by the energized electromagnet 23 transforms the magnetorheological fluid filling the deformation chamber 55 into a near-solid state, further securing the detection base 5 to the detection point. Under the influence of the magnetic field generated by the energized electromagnet 23, a rigid lock is formed between the slider 3 and the groove 21, and between the docking rod 52 and the docking groove 57, ensuring reliable mechanical connections. This eliminates the need for welding or bolts, significantly improving construction efficiency and resolving the problems of poor adaptability and loosening associated with traditional sensors. It also enhances the accuracy of monitoring data and system stability. Furthermore, multiple sensors are incorporated for collaborative monitoring of various physical quantities. Through spatial integration and temporal synchronization of vibration, temperature, noise, and current sensors, a holographic monitoring network for the drive chain is constructed, enabling real-time monitoring of the drive chain's health status and prediction of potential failures.

[0045] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0046] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A wind turbine drivetrain anomaly detection device, comprising a wind turbine drivetrain and detection components, characterized in that: The detection assembly includes a positioning base (2), a detection base (5), a current sensor (7), and an assembly sensor assembly (8). The detection base (5) and the current sensor (7) are movably connected to the positioning base (2). The detection base (5) is fixedly connected to the wind turbine drive chain. The assembly sensor assembly (8) is movably connected to the detection base (5). The assembly sensor assembly (8) includes, but is not limited to, a triaxial vibration sensor, a temperature sensor, and a noise sensor. An electromagnet (23) is embedded in the bottom surface of the positioning base (2). The detection base (5) includes multiple sensor mounting bases (51), and a docking rod (52) is movably connected between two adjacent sensor mounting bases (51). A deformation bladder (55) is connected to the bottom surface of the sensor mounting base (51), and the deformation bladder (55) is filled with magnetorheological fluid. A conical fixing hole (56) is opened in the middle part of the sensor mounting base (51), and a conical fixing sleeve (54) is movably installed in the conical fixing hole (56). The positioning base (2) has sliding grooves (21) on both sides. Multiple sliders (3) are movably connected to the positioning base (2). The sensor fixing seat (51) has docking grooves (57) at both ends. Multiple self-locking pins (6) are embedded in the bottom surface of the sliding groove (21) and the inner wall of the docking groove (57). The self-locking pins (6) are filled with magnetorheological fluid.

2. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: The wind turbine transmission chain includes a turbine housing (1), a generator (11) is fixedly connected inside the turbine housing (1), a bushing (12) is provided on the motor shaft sleeve of the generator (11), a gearbox shaft sleeve is provided on the coaxial line connected to the motor shaft of the generator (11), and a bearing (13) is provided on the gearbox shaft sleeve. The detection base (5) is fixedly connected to the bushing (12) and the bearing (13) respectively. A yaw motor is also connected inside the turbine housing (1), and the lines of the generator (11) and the yaw motor pass through current sensors (7) respectively.

3. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: The bottom surface of the positioning base (2) is also embedded with a permanent magnet (22). The two sides of the slider (3) are fixedly connected with limit blocks (31). The limit blocks (31) are movably locked in the slide groove (21). The bottom surface of the limit blocks (31) is provided with positioning slots (32) arranged at equal intervals. The multiple self-locking pins (6) embedded in the bottom surface of the slide groove (21) are distributed at equal intervals. The distance between two adjacent positioning slots (32) is equal to the distance between adjacent self-locking pins (6) embedded in the bottom surface of the slide groove (21). The bottom surfaces of the limit blocks (31) are provided with arc-shaped end faces (33).

4. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: A self-locking support rod (4) is connected between the detection base (5) and the slider (3). The self-locking support rod (4) includes an outer tube (41) and a telescopic inner rod (42). The bottom end of the telescopic inner rod (42) is movably inserted into the outer tube (41), and a plurality of retaining rings (43) are fixedly connected to one end of the telescopic inner rod (42) located in the outer tube (41). A support spring (44) is provided in the outer tube (41). The support spring (44) is compressed and presses against the bottom of the telescopic inner rod (42) in the initial state. The outer tube (41) is also filled with magnetorheological fluid.

5. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: The sensor mounting base (51) has an inner groove (53) in the middle of its bottom surface. The conical fixing sleeve (54) is inserted into the conical fixing hole (56) from the inner groove (53). The sleeve body of the conical fixing sleeve (54) has a cross-shaped groove (512).

6. The wind turbine drivetrain anomaly detection device according to claim 5, characterized in that: Two docking rods (52) are connected between two adjacent sensor mounting bases (51). One of the docking rods (52) is connected to a ball socket (510), and the other docking rod (52) is connected to a universal ball (511). The universal ball (511) is movably embedded in the ball socket (510).

7. The wind turbine drivetrain anomaly detection device according to claim 6, characterized in that: The inner wall of the docking groove (57) is symmetrically provided with multiple sets of self-locking pins (6). The docking rod (52) is inserted into the docking groove (57) and has an annular docking slot (58) on one end of the rod. The annular docking slot (58) is engaged with one of the sets of self-locking pins (6). The end of the docking rod (52) inserted into the docking groove (57) is provided with a conical end (59).

8. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: The self-locking pin (6) includes a pin housing (61) and a pin inner rod (62). The rod body of the pin inner rod (62) is movably inserted into the pin housing (61). An annular groove (64) is provided on the rod body of the pin inner rod (62) located inside the pin housing (61). A sealing ring (65) is fitted inside the annular groove (64). An arc-shaped end (63) is provided on the outer end of the pin inner rod (62). A locking spring (67) is provided inside the pin housing (61), and the locking spring (67) is compressed in the initial state and presses against the pin inner rod (62) directly below.

9. The wind turbine drivetrain anomaly detection device according to claim 8, characterized in that: The pin housing (61) has a buffer cavity (68) on its side wall. The inner cavity of the pin housing (61) and the buffer cavity (68) are connected by a number of guide holes (69), and the guide holes (69) are located at the bottom of the side wall of the pin housing (61).

10. The wind turbine drivetrain anomaly detection device according to claim 1, characterized in that: The current sensor (7) and each sensor in the assembled sensor assembly (8) are connected to the data acquisition device of the peripheral device. The data acquisition device collects the high-frequency data from the current sensor (7) and each sensor in the assembled sensor assembly (8) and performs preliminary filtering. Then, the data is sent to the edge computing terminal. The edge terminal has a built-in high-speed computing chip and storage unit, and runs spectrum analysis and machine learning algorithms in real time to perform real-time calculation and storage processing on the data. The spectrum analysis and machine learning algorithms include: Offline benchmark model training: Using selected equipment health status data, a benchmark model is trained that includes the target quantity estimate and its reference fluctuation range calculation. Online benchmark model prediction and health assessment utilizes online data of various factors to continuously calculate the estimated values ​​of target quantities and compare them with the measured values ​​to assess the deviation between the two. Furthermore, based on the deviation between the measured values ​​and the dynamic range given by the benchmark model, the health status of the wind turbine is assessed. When the deviation is significant to a certain extent, it is considered that the wind turbine may be in an abnormal state, thereby issuing an early warning.

Images