Tunable device and fault diagnosis method

Abstract

The invention belongs to the technical field of fault signal diagnosis, and particularly relates to a tunable device and a fault diagnosis method. The tunable device comprises a supporting frame with a containing cavity, an adjusting piece connected with the supporting frame, a first magnet located in the containing cavity, a second magnet which is arranged in the containing cavity in a suspended mode and is in magnetic repulsion with the first magnet and an elastic piece with elastic restoring force, and the adjusting piece extends towards the interior of the containing cavity and is connected with the first magnet. One end of the elastic piece is connected with the second magnet, the other end of the elastic piece is connected with the inner wall of the containing cavity, and the adjusting piece is used for adjusting the relative distance between the first magnet and the second magnet so that the second magnet and the to-be-tested device can resonate. According to the invention, the fault characteristic signal can be enhanced, the fault characteristic signal with a high signal-to-noise ratio can be obtained, and the tunable device is simple in structure and low in cost.

Description

Technical Field

[0001] The invention belongs to the technical field of fault signal diagnosis, and in particular relates to a tunable device and a fault diagnosis method. Background Art

[0002] High-end mechanical equipment is an indispensable key component in modern industrial production. Whether it can operate stably is directly related to production efficiency and personnel safety. Various key components in such equipment, including but not limited to rolling bearings, gears, motors, etc., may fail due to long-term operation and complex working conditions. Timely and accurate diagnosis of faults in the operation of mechanical equipment is of great significance to ensure the safe operation of mechanical equipment.

[0003] With the rapid development of computer hardware and sensor technology, intelligent fault diagnosis technology has emerged. This technology deeply integrates traditional methods with computer technology, and realizes the automatic and intelligent diagnosis of mechanical equipment failures through three steps: signal acquisition, fault feature extraction and fault classification. As a powerful feature extraction and classification tool, deep learning shows great potential in fault diagnosis.

[0004] However, traditional deep network models have poor generalization performance and weak robustness when processing complex signals. In order to improve diagnostic performance, researchers continue to explore new deep learning architectures and algorithms, such as introducing feature dimension reduction, attention mechanism, multi-scale feature extraction and other strategies, in order to improve the learning ability and adaptability of the model. Although these improved strategies have improved the diagnostic effect to a certain extent, these algorithms are computationally intensive due to the need for search operations across multiple frequency bands and the requirement for noise reduction. In addition, according to the sampling theorem, the effectiveness of these complex algorithms depends on the ability of the sensing system to maintain a high sampling rate to ensure the capture of effective high-frequency information. In fact, the high-cost sensing system, the large amount of data with high sampling rates, and the complex algorithm deployment are in sharp contradiction with the needs of online monitoring, fault diagnosis, and edge computing in real industrial applications.

[0005] Therefore, there is an urgent need to find a cost-effective method to obtain fault characteristic signals with high signal-to-noise ratio to overcome the detection limitations of existing sensing systems and their related complex algorithms. Summary of the invention

[0006] The purpose of the embodiments of the present application is to provide a tunable device, aiming to solve the problem of how to enhance the fault characteristic signal of the device under test to obtain a fault characteristic signal with a high signal-to-noise ratio and reduce the cost of the device.

[0007] To achieve the above purpose, the technical solution adopted in this application is: In a first aspect, a tunable device is provided for acquiring a vibration signal of a device under test, the tunable device comprising: a supporting frame having a accommodating cavity, an adjusting member connected to the supporting frame, a first magnet located in the accommodating cavity, a second magnet suspended in the accommodating cavity and magnetically repelling the first magnet, and an elastic member having an elastic restoring force, the adjusting member extending into the accommodating cavity and connected to the first magnet, one end of the elastic member being connected to the second magnet, and the other end of the elastic member being connected to an inner wall of the accommodating cavity, the adjusting member being used to adjust a relative distance between the first magnet and the second magnet so that the second magnet resonates with the device under test.

[0008] In some embodiments, an adjustment hole communicating with the accommodating cavity is formed on the surface of the support frame, and one end of the adjustment member is slidably disposed in the adjustment hole.

[0009] In some embodiments, the hole wall of the adjustment hole is provided with an internal thread, the surface of the adjustment member is provided with an external thread matching the internal thread, and the adjustment member is screwed into the adjustment hole.

[0010] In some embodiments, the elastic member is a tube spring.

[0011] In some embodiments, the elastic member is a spiral spring, which is arranged in a planar spiral shape and located at the bottom of the accommodating cavity, the inner end of the spiral spring is connected to the second magnet, and the outer end of the spiral spring is connected to the inner wall of the accommodating cavity.

[0012] In some embodiments, a plurality of the spiral springs are arranged, and any two of the spiral springs are arranged in a multi-turn spiral around the bottom of the accommodating cavity, and the inner ends of the spiral springs are arranged at intervals around the circumference of the second magnet.

[0013] In some embodiments, the tunable device further includes a fixing base having a fixing hole, the elastic member is connected to the fixing base, and the second magnet is located in the fixing hole.

[0014] In some embodiments, the support frame includes a bottom plate connected to the device under test, a top plate arranged relative to the bottom plate, and an annular side plate located between the bottom plate and the top plate, the annular side plate, the bottom plate and the top plate together form the accommodating cavity, the adjusting member is connected to the top plate, and the second magnet is located on the bottom plate when it is in a stationary state.

[0015] In some embodiments, an observation window communicating with the accommodating cavity is provided on the annular side plate.

[0016] In a second aspect, a fault diagnosis method is provided, which uses the tunable device, and the fault diagnosis method comprises the following steps: S1: placing the tunable device on the device under test, and adjusting the distance between the first magnet and the second magnet by the adjusting member so that the second magnet resonates with the device under test; S2: using an acceleration sensor to obtain an acceleration signal of the second magnet, and preprocessing the acceleration signal to obtain an inputtable diagnostic data set; S3: dividing the diagnostic data set into a training set, a validation set and a test set; S4: Build a neural network model and initialize network layer parameters, use a training set to train the neural network model, and use a validation set to test the trained neural network model; determine the neural network model based on whether the trained neural network model meets the preset requirements; S5: Testing the trained neural network model using the test set and outputting the fault diagnosis result.

[0017] The beneficial effect of the present application is that by fixing the tunable device on the device to be tested, under external excitation, after the tunable device is excited by the acceleration of the device to be tested, the second magnet can achieve balance in the upper and lower positions to form a bistable reed-coupled resonance system. The second magnet will generate a maximum vibration displacement corresponding to the fault characteristic signal according to the resonance effect to enhance the fault characteristic signal, and a fault characteristic signal with a high signal-to-noise ratio can be obtained, and the tunable device has a simple structure and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or exemplary technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 is a schematic diagram of the three-dimensional structure of a tunable device provided in an embodiment of the present application; Figure 2 yes Figure 1 A cross-sectional schematic diagram of a tunable device; Figure 3 is an exploded schematic diagram of a tunable device provided in another embodiment of the present application; Figure 4 It is a flowchart of a fault diagnosis method provided by another embodiment of the present application.

[0020] Among them, the reference numerals in the figure are: 100, tunable device; 10, support frame; 20, adjustment member; 31, first magnet; 32, second magnet; 40, elastic member; 111, accommodating cavity; 11, top plate; 13, annular side plate; 131, observation window; 12, bottom plate; 112, adjustment hole; 113, annular groove; 14, guide column; 321, guide hole; DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0022] It should be noted that when a component is referred to as being "fixed on" or "disposed on" another component, it may be directly on the other component or indirectly on the other component. When a component is referred to as being "connected to" another component, it may be directly or indirectly connected to the other component. The orientation or positional relationship indicated by the terms "upper", "lower", "left", "right", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances. The terms "first" and "second" are only used for the purpose of convenience of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features. The meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0023] See also Figures 1 to 3 , an embodiment of the present application provides a tunable device 100 for obtaining a vibration signal of a device under test. The device under test may be a high-end mechanical device, such as a CNC machine tool, an etching machine, a photolithography machine, and an aircraft engine. The device under test can generate vibrations of a certain frequency when in operation. By fixing the tunable device 100 on the device under test, the tunable device 100 can obtain the vibration signal of the device under test, so as to obtain a fault characteristic signal with a high signal-to-noise ratio, and facilitate subsequent fault diagnosis and analysis.

[0024] See also Figures 1 to 3, the tunable device 100 includes: a support frame 10 with a receiving cavity 111, an adjusting member 20 connected to the support frame 10, a first magnet 31 located in the receiving cavity 111, a second magnet 32 ​​suspended in the receiving cavity 111 and magnetically repelling the first magnet 31, and an elastic member 40 with elastic restoring force. It can be understood that the support frame 10 is fixed to the device under test and can vibrate under the drive of the device under test. For the sake of convenience, in the embodiment of the present application, the device under test drives the support frame 10 to vibrate in the vertical direction, that is, the support frame 10 is reciprocating up and down.

[0025] The adjusting member 20 extends into the accommodating cavity 111 and is connected to the first magnet 31, and one end of the elastic member 40 is connected to the second magnet 32 ​​to provide support for the second magnet 32 ​​in a suspended state. It is understandable that the adjusting member 20 can be in the shape of a long strip, and the length direction of the adjusting member 20 is arranged along the vertical direction, and the two ends of the adjusting member 20 are respectively connected to the inner wall of the accommodating cavity 111 and the first magnet 31, and the adjusting member 20 and the first magnet 31 can be connected by a welding process, or the material of the adjusting member 20 is metal iron, so that the adjusting member 20 can be magnetically connected to the first magnet 31, and the selection can be made according to the actual situation, and there is no limitation here.

[0026] See also Figures 1 to 3 , the other end of the elastic member 40 is connected to the inner wall of the accommodating cavity 111. It can be understood that when the support frame 10 vibrates up and down, the second magnet 32 ​​can be driven to vibrate in the vertical direction, and the second magnet 32 ​​will stretch or compress the elastic member 40 during the movement. The adjusting member 20 is used to adjust the relative distance between the first magnet 31 and the second magnet 32, that is, when the second magnet 32 ​​is in a stationary state, the adjusting member 20 can drive the first magnet 31 to approach the second magnet 32, or drive the first magnet 31 away from the second magnet 32, thereby adjusting the distance between the first magnet 31 and the second magnet 32 ​​in the stationary state, and then adjusting the magnetic attraction between the first magnet 31 and the second magnet 32 ​​during the vibration process, and finally realizing the adjustment of the resonant frequency of the tunable device 100. The resonant frequency of the tunable device 100 is made equal to the resonant frequency of the fault characteristic signal of the device under test, so that the tunable device 100 can resonate with the device under test, that is, resonate, so that the second magnet 32 ​​can provide the maximum vibration displacement output, thereby enhancing the fault characteristic signal.

[0027] See also Figures 1 to 3In the embodiment of the present application, by fixing the tunable device 100 on the device to be tested, under external excitation, after the tunable device 100 is excited by the acceleration of the device to be tested, the second magnet 32 ​​can achieve balance at the upper and lower positions to form a bistable reed-coupled resonance system. The second magnet 32 ​​will generate a maximum vibration displacement corresponding to the fault characteristic signal according to the resonance effect to enhance the fault characteristic signal, and a fault characteristic signal with a high signal-to-noise ratio can be obtained. In addition, the tunable device 100 has a simple structure and low cost.

[0028] See also Figures 1 to 3 It can be understood that the resonant frequency corresponding to different fault characteristic signals can also be satisfied by adjusting the position of the first magnet 31, that is, according to different fault characteristic signals, the first magnet 31 is placed in different positions, so that the first magnet 31 can produce a large displacement response output under different fault characteristics, thereby enhancing the fault characteristic signal and improving the application range of the tunable device 100. A single device can meet the enhancement of multiple different fault characteristic signals and reduce costs.

[0029] Optionally, the acquired fault feature signal can be input into a deep learning network for diagnosis to achieve a low sampling rate, thereby completing the detection and diagnosis of the fault without the need for complex algorithm deployment.

[0030] See also Figures 1 to 3 Optionally, the first magnet 31 is located directly above the second magnet 32, and the second magnet 32 ​​vibrates relative to the first magnet 31. The first magnet 31 and the second magnet 32 ​​have negative stiffness, and the spring has positive stiffness. The stiffness of the tunable device 100 is equal to the positive stiffness of the spring plus the negative stiffness of the first magnet 31 and the second magnet 32. By adjusting the distance between the first magnet 31 and the second magnet 32, the negative stiffness value of the first magnet 31 and the second magnet 32 ​​can be adjusted, thereby achieving overall stiffness adjustment of the tunable device 100.

[0031] Optionally, a magnetic field can be generated between the first magnet 31 and the second magnet 32 ​​. The first magnet 31 and the second magnet 32 ​​can be permanent magnets, static magnets or constant magnets, which are not limited here and can be selected according to actual conditions.

[0032] Optionally, the first magnet 31 and the second magnet 32 ​​may both be electromagnets, so that the magnetic attraction between the first magnet 31 and the second magnet 32 ​​may be changed by changing the magnitude of the current, that is, the negative stiffness value between the first magnet 31 and the second magnet 32 ​​may be changed.

[0033] See also Figures 1 to 3Optionally, the first magnet 31 is a magnet, and the second magnet 32 ​​is an electromagnet. By changing the position of the first magnet 31, the negative stiffness value between the first magnet 31 and the second magnet 32 ​​is changed, and the magnitude of the current of the second magnet 32 ​​is changed, and the negative stiffness value between the first magnet 31 and the second magnet 32 ​​can also be changed, so as to achieve the adjustment of the resonant frequency of the tunable device 100. It can be understood that the adjustment accuracy of the position of the first magnet 31 is lower than the adjustment accuracy of the current of the second magnet 32, that is, the adjustment member 20 can first adjust the position of the first magnet 31 to achieve the first-level rough adjustment, and then achieve the second-level fine adjustment by adjusting the current of the second magnet 32, and finally achieve the precise adjustment of the overall resonant frequency of the tunable device 100.

[0034] Optionally, the first magnet 31 includes a magnet and a coil disposed on the magnet, and the second magnet 32 ​​also includes a magnet and a coil disposed on the magnet. For example, the magnet is ring-shaped, and the coil is wound around the circumference of the magnet, so that the negative stiffness value between the first magnet 31 and the second magnet 32 ​​can be adjusted by changing the position of the first magnet 31 and / or changing the current of the coil.

[0035] It is understandable that the use of electromagnets can improve the signal-to-noise ratio and reduce external interference, thereby improving the accuracy of subsequent fault diagnosis.

[0036] See also Figures 1 to 3 In some embodiments, an adjustment hole 112 communicating with the accommodating cavity 111 is opened on the surface of the support frame 10 , and one end of the adjustment member 20 is slidably disposed in the adjustment hole 112 .

[0037] Optionally, the other end of the adjusting member 20 extends into the accommodating cavity 111 and is connected to the first magnet 31. The adjusting member 20 may be a screw. After the adjusting member 20 slides into place, two nuts are screwed on the adjusting member 20, one of which is located in the accommodating cavity 111 and the other is located outside the accommodating cavity 111, so that the adjusting member 20 can be fixed on the supporting frame 10 to achieve the adjustment of the resonant frequency of the tunable device 100.

[0038] See also Figures 1 to 3 In some embodiments, the wall of the adjusting hole 112 is provided with an internal thread, the surface of the adjusting member 20 is provided with an external thread matching the internal thread, and the adjusting member 20 is screwed into the adjusting hole 112 .

[0039] Optionally, by rotating the adjusting member 20 relative to the supporting frame 10, the length of the adjusting member 20 in the accommodating cavity 111 can be adjusted, so that the distance between the first magnet 31 and the second magnet 32 ​​in a static state can be adjusted. Through the cooperation of the internal thread and the external thread, the adjustment process can be simplified, and precise adjustment can be achieved to avoid accidental sliding or loosening during the adjustment process.

[0040] In some embodiments, the elastic member 40 is a tube spring.

[0041] Optionally, the length direction of the tube spring is arranged in the vertical direction, one end of the tube spring is connected to the bottom of the accommodating cavity 111, the adjusting member 20 is connected to the top of the accommodating cavity 111, the tube spring is located directly below the first magnet 31, and the second magnet 32 ​​is located at the upper end of the tube spring. The second magnet 32 ​​will compress and stretch the tube spring during the vibration process, thereby forming a bistable nonlinear dynamic model. The second magnet 32 ​​can achieve a larger displacement output. Through the resonance principle, the signal amplified by the weak fault feature in the external excitation can be detected in the displacement output at the second magnet 32, and then the fault feature signal is input into the deep learning algorithm for fault diagnosis.

[0042] See also Figures 1 to 3 In some embodiments, the elastic member 40 is a spiral spring, which is arranged in a planar spiral shape and is located at the bottom of the accommodating cavity 111 , the inner end of the spiral spring is connected to the second magnet 32 ​​, and the outer end of the spiral spring is connected to the inner wall of the accommodating cavity 111 .

[0043] Optionally, the spiral spring is at a certain distance from the bottom of the accommodating cavity 111, the spiral spring is in a cantilever beam type, and the second magnet 32 ​​is located at the free end of the spiral spring. The spiral spring can effectively reduce the space occupied by the accommodating cavity 111, which is conducive to the miniaturization of the volume of the tunable device 100.

[0044] Optionally, the spiral spring may be made of metal, such as stainless steel, and the support frame 10 may be made of metal, such as iron. The outer end of the spiral spring may be fixedly connected to the cavity wall of the accommodating cavity 111 by welding.

[0045] Optionally, an inner wall of the accommodating cavity 111 is provided with an annular groove 113 , and an edge of an outer end of the spiral spring is disposed in the annular groove 113 .

[0046] See also Figures 1 to 3 In some embodiments, multiple spiral springs are arranged, and any two spiral springs are arranged in a multi-turn spiral around the bottom of the accommodating cavity 111, and the inner ends of each spiral spring are arranged at intervals around the circumference of the second magnet 32.

[0047] Optionally, the three spiral springs are arranged in the same plane in a static state. During the vibration process, the three spiral springs are stretched and deformed, and act together on the second magnet 32, so that the second magnet 32 ​​is evenly stressed. The second magnet 32 ​​can vibrate stably in the vertical direction, and by arranging the spiral springs in a surrounding manner, the load can be distributed more evenly, reducing the fatigue of a single elastic member 40, thereby improving the service life of the entire device and the stability of fault diagnosis.

[0048] It can also be understood that any two spiral springs are arranged in multiple turns of spiral around the bottom of the accommodating cavity 111, and more spiral springs can be arranged within a unit area, so that multiple parallel spiral springs can act together on the second magnet 32 ​​and limit the displacement of the second magnet 32 ​​relative to the vertical direction, which is beneficial to the resonance effect of the second magnet 32 ​​along the vertical direction and is also beneficial to the compactness of the overall structure of the tunable device 100.

[0049] Optionally, a guide hole 321 is opened at the center position of the second magnet 32, and a guide column 14 is set at the bottom of the accommodating cavity 111. The guide column 14 is slidably inserted into the guide hole 321, so that through the cooperation of the guide hole 321 and the guide column 14, the second magnet 32 ​​can be guided to slide in the vertical direction to avoid the second magnet 32 ​​from shifting during vibration.

[0050] See also Figures 1 to 3 In some embodiments, the tunable device 100 further includes a fixing base having a fixing hole, the elastic member 40 is connected to the fixing base, and the second magnet 32 ​​is located in the fixing hole.

[0051] Optionally, the second magnet can be fixed by a fixing seat, and the elastic member 40 and the fixing seat can be connected by a welding process without destroying the structural characteristics of the second magnet 32. It can also ensure that the second magnet 32 ​​does not shift or misalign during vibration, thereby maintaining the resonance characteristics of the tunable device 100.

[0052] See also Figures 1 to 3 In some embodiments, the support frame 10 includes a bottom plate 12 connected to the device under test, a top plate 11 arranged relative to the bottom plate 12, and an annular side plate 13 located between the bottom plate 12 and the top plate 11. The annular side plate 13, the bottom plate 12 and the top plate 11 together form a accommodating cavity 111. The adjusting member 20 is connected to the top plate 11. When the second magnet 32 ​​is in a stationary state, it is located on the bottom plate 12.

[0053] Optionally, the bottom plate 12 can be fixed to the device to be tested by a clamp, and the bottom plate 12, the top plate 11 and the annular side plate 13 can ensure a reasonable spatial layout of the accommodating cavity 111. The cross-sectional shape of the accommodating cavity 111 is circular, so that the three spiral springs can be arranged symmetrically, so that the first magnet 31, the second magnet 32 ​​and the elastic member 40 can work together in the accommodating cavity 111.

[0054] See also Figures 1 to 3 In some embodiments, an observation window 131 communicating with the accommodating cavity 111 is formed on the annular side plate 13 .

[0055] Optionally, the observation window 131 allows the user to observe the working status inside the accommodating cavity 111 during operation, thereby improving the convenience of operating the device. The sliding status of the second magnet 32, the restoring force of the elastic member 40 and the working condition of the adjusting member 20 can be directly viewed through the observation window 131, thereby timely adjusting and optimizing the working status of the device and improving the efficiency and accuracy of fault diagnosis.

[0056] Optionally, the observation window 131 can also reduce the mass of the support frame 10. The reduction in the mass of the support frame 10 helps to improve the response speed and sensitivity of the system, so that the resonance frequency can better match the fault characteristic signal frequency of the device under test, thereby facilitating resonance. It is understandable that the smaller the mass of the support frame 10, the smaller its own inertia. A system with small inertia responds faster to external vibrations (such as vibrations of the device under test), can more sensitively follow the vibrations of the device under test, and thus more easily reaches a resonant state.

[0057] See also Figure 4 The present invention also proposes a fault diagnosis method, which uses the above-mentioned tunable device 100. The specific structure of the tunable device 100 refers to the above-mentioned embodiment. Since the fault diagnosis method adopts all the technical solutions of all the above-mentioned embodiments, it also has all the beneficial effects brought by the technical solutions of the above-mentioned embodiments, which will not be repeated here one by one.

[0058] The fault diagnosis method includes the following steps: S1: placing the tunable device 100 on the device under test, and adjusting the distance between the first magnet 31 and the second magnet 32 ​​by the adjusting member 20 so that the second magnet 32 ​​resonates with the device under test; S2: Use an acceleration sensor to obtain the acceleration signal of the second magnet 32, and preprocess the acceleration signal to obtain an input diagnostic data set; it can be understood that the acceleration sensor is connected to the second magnet 32 ​​to obtain the acceleration signal of the second magnet 32 ​​during the vibration process, and the acceleration signal contains the fault characteristic signal of the device to be tested.

[0059] S3: Divide the diagnostic dataset into training set, validation set and test set; S4: Build a neural network model and initialize the network layer parameters, use the training set to train the neural network model, and use the validation set to test the trained neural network model; determine the neural network model based on whether the trained neural network model meets the preset requirements; S5: Test the trained neural network model using the test set and output the fault diagnosis results.

[0060] The fault diagnosis method provided by the present application enhances the fault characteristic signal in the mechanical operation equipment, and can also adapt to the frequency of different fault characteristics and the signal of the working condition by adjusting the tunable device 100 to be located at different positions of the mechanical device, and the measured fault characteristic signal is used to realize efficient online fault diagnosis through a deep learning algorithm, thereby realizing a new combination of mechanical and computer algorithms. Compared with the existing single fault diagnosis method, the scheme of the present application utilizes the mechanical dynamic resonance characteristics combined with a deep learning algorithm to strive to solve the limitations of traditional methods and sensor systems. It includes reducing the data sampling rate, simplifying the deployment process of the demodulation algorithm, and improving the local sensitivity of the sensor. At the same time, a lower sampling rate, that is, a lower amount of data, can be used to enable the deep learning network to learn fault characteristics and achieve efficient fault diagnosis.

[0061] See also Figure 4 Through this method, the user can use the tunable device 100 in conjunction with the device under test, accurately adjust the relative position of the second magnet 32, and make it resonate with the device under test, thereby improving the response accuracy to the fault characteristic signal. The diagnostic data set formed by the preprocessing of the signal collected by the acceleration sensor can be input into the neural network for training and testing, ensuring accurate diagnosis of the fault of the device under test through intelligent analysis.

[0062] See also Figure 4 The method of this application combines hardware and artificial intelligence technology, improves the automation level and accuracy of fault diagnosis, provides a scientific basis for equipment maintenance, and can greatly improve the stability and life of equipment operation.

[0063] See also Figure 4Furthermore, the tunable device 100 provided in the embodiment of the present application solves the problem of single method in the field of fault diagnosis from the perspective of actual industrial application, and broadens the efficient online diagnosis of different fault conditions. The method is based on the most widely used vibration signal. By attaching the tunable device 100 to the surface of the device to be tested, the device can provide a resonant frequency band that can be artificially generated. The fault characteristic signal of the device to be tested is enhanced by the resonant frequency band of the device to obtain a fault characteristic signal with a high signal-to-noise ratio. By inputting the enhanced fault characteristic signal into the deep learning network for fault diagnosis, the limitations of traditional methods and sensor systems are solved. This fault diagnosis method can not only reduce the data sampling rate and simplify the deployment process of the demodulation algorithm, but also improve the local sensitivity of the sensor to achieve efficient fault diagnosis.

[0064] The above are only optional embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A tunable device for obtaining a vibration signal of a device under test, characterized in that: The tunable device includes: a supporting frame with a receiving cavity, an adjusting member connected to the supporting frame, a first magnet located in the receiving cavity, a second magnet suspended in the receiving cavity and magnetically repelling the first magnet, and an elastic member with elastic restoring force, wherein the adjusting member extends into the receiving cavity and is connected to the first magnet, one end of the elastic member is connected to the second magnet, and the other end of the elastic member is connected to the inner wall of the receiving cavity, and the adjusting member is used to adjust the relative distance between the first magnet and the second magnet so that the second magnet resonates with the device under test.

2. The tunable device according to claim 1, characterized in that: An adjusting hole communicating with the accommodating cavity is provided on the surface of the supporting frame, and one end of the adjusting member is slidably disposed in the adjusting hole.

3. The tunable device according to claim 2, characterized in that: The hole wall of the adjusting hole is provided with an internal thread, the surface of the adjusting member is provided with an external thread matching the internal thread, and the adjusting member is screwed in the adjusting hole.

4. The tunable device according to any one of claims 1 to 3, characterized in that: The elastic member is a tube spring.

5. The tunable device according to any one of claims 1 to 3, characterized in that: The elastic member is a spiral spring, which is arranged in a planar spiral shape and is located at the bottom of the accommodating cavity. The inner end of the spiral spring is connected to the second magnet, and the outer end of the spiral spring is connected to the inner wall of the accommodating cavity.

6. The tunable device according to claim 5, characterized in that: A plurality of the spiral springs are arranged, and any two of the spiral springs are arranged in a multi-turn spiral around the bottom of the accommodating cavity, and the inner ends of the spiral springs are arranged at intervals in the circumferential direction around the second magnet.

7. The tunable device according to any one of claims 1 to 3, characterized in that: The tunable device also includes a fixing seat having a fixing hole, the elastic member is connected to the fixing seat, and the second magnet is located in the fixing hole.

8. The tunable device according to any one of claims 1 to 3, characterized in that: The support frame includes a bottom plate connected to the device under test, a top plate arranged relative to the bottom plate, and an annular side plate located between the bottom plate and the top plate. The annular side plate, the bottom plate and the top plate together form the accommodating cavity. The adjusting member is connected to the top plate. When the second magnet is in a stationary state, it is located on the bottom plate.

9. The tunable device according to claim 8, characterized in that: The annular side plate is provided with an observation window connected to the accommodating cavity.

10. A fault diagnosis method, using the tunable device according to any one of claims 1 to 9, characterized in that: The fault diagnosis method comprises the following steps: S1: placing the tunable device on the device under test, and adjusting the distance between the first magnet and the second magnet by the adjusting member so that the second magnet resonates with the device under test; S2: using an acceleration sensor to obtain an acceleration signal of the second magnet, and preprocessing the acceleration signal to obtain an inputtable diagnostic data set; S3: dividing the diagnostic data set into a training set, a validation set and a test set; S4: Build a neural network model and initialize network layer parameters, use a training set to train the neural network model, and use a validation set to test the trained neural network model; determine the neural network model based on whether the trained neural network model meets the preset requirements; S5: Testing the trained neural network model using the test set and outputting the fault diagnosis result.

Images