Air bearing turbine wheel detection and identification device

Abstract

This invention discloses an air bearing turbine impeller detection and identification device, including a base and a control system. The base has an air-bearing vibration isolation unit at its bottom, and a double-hemispherical air bearing assembly mounted on its upper surface. A quick-change clamp is connected to the front end of the main shaft of the double-hemispherical air bearing assembly. A non-contact drive assembly for rotating the suspended turbine impeller is mounted on one side of the double-hemispherical air bearing assembly. A segmented detection ring is slidably mounted on the base, capable of horizontal movement to cover or avoid the turbine impeller. The segmented detection ring is equipped with a set of intelligent sensors for detection, and an acoustic emission sensor set is embedded inside the double-hemispherical air bearing assembly. This invention uses a double-hemispherical air bearing assembly as the core support for the impeller main shaft, ensuring that the impeller remains in a pure air-film suspension state without mechanical contact during the detection process, eliminating the frictional noise, vibration interference, and additional imbalance caused by traditional mechanical bearing supports.

Description

Technical Field

[0001] This invention relates to the field of detection and fault diagnosis technology, and in particular to an air bearing turbine impeller detection and identification device. Background Technology

[0002] Turbine impellers (such as core components of aero engines, gas turbines, and turbochargers) are core power components of high-speed rotating machinery. Their manufacturing precision, dynamic balance performance, and structural integrity directly determine the overall efficiency, lifespan, and operational safety of the machine. As advanced manufacturing technologies develop towards higher precision, lighter weight, and higher performance, the testing requirements for turbine impellers have evolved from simple static geometric measurements to dynamic comprehensive testing under simulated real-world conditions, involving multiple physical fields, all dimensions, and intelligent operation.

[0003] The current common practice is to detect imbalance and vibration on a dynamic balancing machine supported by mechanical bearings; acquire geometric parameters such as blade profile and thickness using a coordinate measuring machine or optical scanner; detect surface cracks and defects using industrial vision equipment or fluorescent flaw detection; and rely on offline ultrasonic or CT inspection for internal defects (such as microcracks and material inclusions). This equipment- and process-specific inspection mode leads to low efficiency, inconsistent spatiotemporal data benchmarks at each stage, and difficulty in comprehensively evaluating and analyzing the impeller's condition, thus failing to meet the demands of high-efficiency, integrated intelligent manufacturing.

[0004] Furthermore, traditional dynamic balancing and dynamic testing equipment generally uses mechanical rolling bearings or sliding bearings as supports. The inherent friction, clearance, and vibration of these bearings directly couple into the test signal, forming "background noise" that severely interferes with the accurate identification of minute imbalances and abnormal vibrations in the impeller body. More importantly, for advanced turbomachinery (such as air circulators and micro gas turbines) that widely use air bearings, the impeller is in a non-contact suspended state during actual operation. The testing environment supported by rigid mechanical bearings differs fundamentally from actual operating conditions, resulting in a weak correlation between the test results and actual operating performance, raising questions about their reliability.

[0005] Furthermore, existing online or integrated inspection equipment mostly focuses on geometric quantities (such as profile and runout) or visible surface defects. For internal damage that may initiate or expand during operation (such as high-cycle fatigue microcracks and internal material defects), there is a lack of effective online, non-contact monitoring methods. While acoustic emission technology can effectively monitor internal material damage, traditional acoustic emission sensors require close coupling to the surface of the workpiece (contact mounting), making them difficult to apply to high-speed rotating impeller inspection scenarios that require non-contact support, thus creating a blind spot in technology application.

[0006] In summary, the existing technology system suffers from a combination of bottlenecks, including distorted detection status, limited detection capabilities, and a lack of internal detection. Summary of the Invention

[0007] The technical problem solved by the present invention is to provide an air bearing turbine impeller detection and identification device.

[0008] This application provides an air bearing turbine impeller detection and identification device, including a base and a control system. The base has an air-bearing vibration isolation unit at its bottom, and a double-hemispherical air bearing assembly for suspending and supporting the turbine impeller's main shaft is installed in the central area of ​​its upper surface. A quick-change clamp for clamping the turbine impeller is connected to the front end of the main shaft of the double-hemispherical air bearing assembly. A non-contact drive assembly for driving the suspended turbine impeller to rotate is installed on one side of the double-hemispherical air bearing assembly. A segmented detection ring is slidably disposed on the base, and the segmented detection ring can move horizontally to cover or avoid the turbine impeller. A smart sensor group for detection is provided on the segmented detection ring, and an acoustic emission sensor group is embedded inside the double-hemispherical air bearing assembly. The smart sensor group, the acoustic emission sensor group, and the drive assembly are all electrically connected to the control system.

[0009] Furthermore, the dual-hemispherical air bearing assembly includes a bearing housing and an upper hemisphere bearing and a lower hemisphere bearing installed within the bearing housing; the bearing housing engages with a positioning stop on the base, and at least two annular grooves are machined on the inner spherical surfaces of both the upper and lower hemisphere bearings, which divide the inner surface of each hemisphere bearing into at least three independent pressure zones; each pressure zone has an independent air intake channel on its back; each air intake channel is externally connected to an electro-proportional valve 47 controlled by the control system, and each pressure zone is correspondingly provided with a displacement sensor for monitoring the impeller journal suspension clearance.

[0010] Furthermore, the acoustic emission sensor array is embedded in the side wall of the bearing housing with an interference fit, and its sensing end face is flush with the inner cavity of the bearing.

[0011] Furthermore, the segmented detection ring includes a lower half-ring fixed base, an upper half-ring movable cover hinged thereto, and a locking mechanism that fixes the two together; the lower half-ring fixed base is provided with an arc-shaped guide rail along the circumferential direction and a standard electrical interface on its inner side; each sensor of the intelligent sensor group is detachably mounted on the arc-shaped guide rail by a slider at its bottom, and its position is fixed by a locking knob; when in place during installation, the electrical plug on the sensor automatically connects to the standard electrical interface.

[0012] Furthermore, the drive assembly is a pneumatic turbine drive ring fixed to the base and located below the turbine impeller, which includes multiple miniature high-speed solenoid valve nozzles that are circumferentially distributed and have adjustable injection direction.

[0013] Furthermore, it also includes a non-contact braking assembly, which includes a brake disc coaxially connected to the main shaft and a brake fixed to the base.

[0014] Furthermore, the quick-change clamp is a hydraulic expansion mandrel, which is fitted with a replaceable modular adapter.

[0015] Furthermore, the smart sensor group includes at least two of the following: a laser displacement sensor, a 3D scanner, and an industrial camera.

[0016] Furthermore, the industrial camera is mounted on a slide that can move along the arc-shaped guide rail. The slide is integrated with a servo drive mechanism and can perform continuous or point-to-point circumferential movement on the arc-shaped guide rail. The air-bearing vibration isolation unit has a honeycomb sandwich structure with interconnected honeycomb-shaped air chambers inside.

[0017] Furthermore, each sensor is used to collect multi-source signals during the rotation of the turbine impeller. The control system fuses the collected multi-source spatiotemporal synchronous data and automatically identifies defects and generates a comprehensive inspection report through a pre-trained AI model. The multi-source signals include vibration signals, three-dimensional point cloud data, surface image data, infrared thermal imaging data, and acoustic emission signals. The AI ​​model is a deep learning model based on the fusion of multi-source signals.

[0018] Compared with existing technologies, this invention uses a double-hemispherical air bearing assembly as the core support for the impeller main shaft, ensuring the impeller remains in a pure air film suspension state without mechanical contact during the inspection process. This eliminates the frictional noise, vibration interference, and additional imbalance caused by traditional mechanical bearing supports. Through the intelligent sensor group (such as a laser displacement sensor and a 3D scanner) on the segmented detection ring, geometric parameters such as impeller profile, blade profile, and radial runout can be acquired online. By directly embedding the acoustic emission sensor group inside the double-hemispherical air bearing assembly, utilizing the air film of the air bearing as a natural acoustic coupling medium, non-contact, online real-time monitoring of microcrack initiation and propagation, material defects, and other damage inside high-speed rotating impellers is achieved for the first time. Furthermore, a combination design of the segmented detection ring and quick-change fixture is employed. The detection ring can open and close horizontally, facilitating vertical impeller hoisting and protecting precision blades. The quick-change fixture supports hydraulic expansion and other methods, enabling rapid switching between impellers of different specifications. The entire changeover process can be completed within minutes, significantly reducing equipment downtime. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0020] Figure 1 This is a schematic diagram of the air bearing turbine impeller detection and identification device of the present invention; Figure 2This is a cross-sectional view of the double hemisphere air bearing assembly of the present invention; Figure 3 This is a top view of the double hemisphere air bearing assembly of the present invention; Figure 4 This is a partial structural schematic diagram of the lower half-ring fixing base of the present invention; Figure 5 This is a side view of the lower half-ring fixing base of the present invention; Figure 6 This is a top view of the segmented detection ring of the present invention; Figure 7 This is a cross-sectional view of the turbine impeller of the present invention; Figure 8 This is a structural block diagram of the air bearing turbine impeller detection and identification device of the present invention.

[0021] The reference numerals in the attached figures include: 1. Base; 2. Control system; 3. Air-bearing vibration isolation unit; 4. Double hemisphere air bearing assembly; 41. Bearing housing; 42. Upper hemisphere bearing; 43. Lower hemisphere bearing; 44. Annular groove; 45. Pressure zone; 46. Inlet passage; 47. Electro-proportional valve; 48. Displacement sensor; 49. Elastic retaining ring; 5. Turbine impeller; 6. Quick-change fixture; 7. Drive assembly; 8. Split detection ring; 81. Lower half ring fixing base; 82. Upper half... 83. Ring-shaped movable cover; 84. Locking mechanism; 85. Slider; 86. Arc-shaped guide rail; 87. Standard electrical interface; 9. Intelligent sensor group; 91. Laser displacement sensor; 92. 3D scanner; 93. Industrial camera; 94. Slide table; 10. Acoustic emission sensor group; 11. Electrical plug; 12. Miniature high-speed solenoid valve nozzle; 13. Braking assembly; 131. Brake disc; 132. Brake; 14. Linear guide rail; 15. Modular adapter. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0023] like Figures 1 to 8As shown, the air bearing turbine impeller detection and identification device of the present invention includes a base 1 and a control system 2. The base 1 is made of high-rigidity cast iron through aging treatment, with precision machining on the upper surface. An air-bearing vibration isolation unit 3 is provided at its bottom. A double-hemispherical air bearing assembly 4 for suspending and supporting the main shaft of the turbine impeller 5 is installed in the central area of ​​its upper surface. The front end of the main shaft of the double-hemispherical air bearing assembly 4 is connected to a quick-change clamp 6 for clamping the turbine impeller 5. A non-contact drive assembly 7 for driving the suspended turbine impeller 5 to rotate is installed on one side of the double-hemispherical air bearing assembly 4. A segmented detection ring 8 is slidably arranged on the base 1. The segmented detection ring 8 can move horizontally to cover or avoid the turbine impeller 5. A smart sensor group 9 for detection is provided on the segmented detection ring 8. An acoustic emission sensor group 10 is embedded inside the double-hemispherical air bearing assembly 4. The smart sensor group 9, the acoustic emission sensor group 10, and the drive assembly 7 are all electrically connected to the control system 2.

[0024] This invention employs a double-hemispherical air bearing assembly 4 as the core support for the impeller's main shaft, ensuring the impeller remains in a pure air-film suspension state without mechanical contact during testing. This eliminates the frictional noise, vibration interference, and additional imbalance caused by traditional mechanical bearing supports. The intelligent sensor group 9 (such as a laser displacement sensor 91 and a 3D scanner 92) on the segmented detection ring 8 can acquire geometric parameters such as impeller profile, blade profile, and radial runout online. By directly embedding the acoustic emission sensor group 10 inside the double-hemispherical air bearing assembly 4, utilizing the air film of the air bearing as a natural acoustic coupling medium, non-contact, online real-time monitoring of microcrack initiation and propagation, material defects, and other damage inside high-speed rotating impellers is achieved for the first time. Furthermore, a combined design of the segmented detection ring 8 and the quick-change fixture 6 is used. The detection ring can open and close horizontally, facilitating vertical impeller hoisting and protecting precision blades. The quick-change fixture 6 supports hydraulic expansion and other methods, enabling rapid switching between impellers of different specifications. The entire changeover process can be completed within minutes, significantly reducing equipment downtime.

[0025] In some embodiments, such as Figure 2 , Figure 3As shown, the double hemisphere air bearing assembly 4 includes a bearing housing 41 and an upper hemisphere bearing 42 and a lower hemisphere bearing 43 installed within the bearing housing 41. The bearing housing 41 is precisely fitted with a circular positioning stop on the base 1 through a protrusion at its lower part and is fastened with screws. Both the upper hemisphere bearing 42 and the lower hemisphere bearing 43 are made of porous graphite, and the two hemispheres are axially fixed by elastic retaining rings 49 through two holes at the top and bottom. On the inner spherical surface of each hemisphere bearing, three annular grooves 44 are precisely machined, thereby dividing the inner surface of each hemisphere into four independent pressure regions 45: A, B, C, and D. Each pressure region 45 has an independent air intake channel 46 on its back, leading to the outside of the bearing housing 41. Each air intake channel 46 is externally connected to an electro-proportional valve 47 controlled by the control system 2, and each pressure region 45 is correspondingly provided with a displacement sensor 48 for monitoring the impeller journal suspension clearance. Its probe is perpendicularly pointed to the journal surface at the center of the four pressure regions 45, for real-time measurement of the air film clearance. The acoustic emission sensor group 10 is embedded in the side wall of the bearing housing 41 with an interference fit. It includes 8 acoustic emission sensors that are evenly distributed around the circumference. The sensing end face is flush with the inner cavity of the bearing and receives signals through the air film coupling of the air bearing.

[0026] In this embodiment, at least three independent pressure regions 45 separated by annular grooves 44 are respectively set on the inner spherical surface of the upper and lower hemispherical bearings 43, and corresponding independent electro-proportional valves 47 and high-precision displacement sensors 48 are configured to form a closed-loop active control system 2 with multiple inputs and multiple outputs. It can monitor the suspension gap of the impeller journal in each region in real time and dynamically and independently adjust the air film pressure in the corresponding region, thereby actively compensating for the initial mass eccentricity of the impeller and realizing automatic centering and leveling in three-dimensional space. This setting not only improves the suspension stability under high-speed rotation, but also ensures the high consistency of the support state after each clamping.

[0027] Furthermore, the acoustic emission sensor assembly 10 is directly embedded into the side wall of the bearing housing 41 with an interference fit, and its sensing end face is flush with the pressure-bearing air film area of ​​the bearing cavity. This design makes the working air film of the air bearing itself a near-ideal sound wave propagation medium. The high-frequency stress wave signal released when the impeller rotates or internal damage occurs can be efficiently transmitted to the embedded sensor through this air film, thus realizing truly non-contact, online, and real-time acoustic emission monitoring. This integrated design allows the device to simultaneously acquire two types of key data in a single rotation test: first, the overall mass distribution and dynamic rigidity information of the impeller reflected by the multi-zone leveling system; and second, the information on microscopic damage activity inside the material captured by the embedded acoustic emission system. The spatiotemporal synchronization and correlation analysis of these two types of information enables a comprehensive and in-depth diagnosis of the impeller from its macroscopic dynamic balance state to its microscopic structural integrity state.

[0028] Specifically, displacement sensors 48 are installed at the upper and lower ends of the bearing housing 41 end cover. The upper sensor is installed at the center of the upper end cover, pointing vertically downwards towards the spindle journal. This measuring point is located above the upper hemispherical bearing 42, and the lower sensor is installed at the center of the lower end cover, pointing vertically upwards towards the spindle journal. This measuring point is located below the lower hemispherical bearing 43. By comparing the gap values ​​measured by the upper and lower sensors, the spindle tilt angle can be accurately calculated. The upper hemispherical bearing 42 has four pressure zones: A1, B1, C1, and D1, while the lower hemispherical bearing has A2, B2, C2, and D2. The data from the upper displacement sensor 48 comprehensively reflects the overall support effect of all pressure zones in the upper hemisphere on the spindle, and is used to adjust the pressure distribution of each zone in the upper hemisphere. The data from the lower displacement sensor 48 correspondingly adjusts the pressure of each zone in the lower hemisphere. The combined data from the upper and lower sensors are used to coordinate the total support force of the upper and lower hemispheres, controlling the spindle's axial "suspended height" and "tilt attitude".

[0029] In some embodiments, such as Figure 4 , Figure 6 As shown, the segmented detection ring 8 includes a lower half-ring fixed base 81, an upper half-ring movable cover 82 hinged thereto, and a locking mechanism 83 (such as an electromagnetic lock, electromagnetic pin, pneumatic latch, etc.) that fixes the two together; the lower half-ring fixed base 81 is provided with an arc-shaped guide rail 85 along the circumferential direction and a standard electrical interface 86 on its inner side; each sensor of the intelligent sensor group 9 is detachably installed on the arc-shaped guide rail 85 by a slider 84 at its bottom, and its position is fixed by a locking knob; when installed in place, the electrical plug 11 on the sensor automatically connects to the standard electrical interface 86.

[0030] In this embodiment, the detection ring is designed as a split structure with the lower half-ring fixed base 81 and the upper half-ring movable cover 82 connected by a hinge. During loading, simply open the upper half-ring and move the entire detection ring horizontally out. This allows for vertical hoisting of the impeller in a completely open and unobstructed space, preventing accidental collisions between the impeller (especially its fragile blades) and the detection device, significantly reducing operational risks and workpiece damage rates. This is particularly suitable for the safe and efficient loading and unloading of large, high-value impellers. Furthermore, an arc-shaped guide rail 85 along the circumferential direction is provided inside the detection ring, and each sensor is designed as an independent module with a standardized slider and electrical plug 11. When replacing a sensor or adjusting its circumferential position, simply loosen the locking knob to slide it along the guide rail or remove / install it entirely. After the sensor is in place, its electrical plug 11 automatically connects to the standard electrical interface 86 inside the ring, achieving plug-and-play functionality. This design allows for sensor layout adjustments and overall detection ring function configuration to be completed within minutes, enabling a single device to quickly adapt to diverse detection tasks.

[0031] Furthermore, after closing and locking, the lower half-ring and the upper half-ring form a complete, closed rigid ring. All sensors are directly mounted and locked onto the arc-shaped guide rail 85 of this rigid ring via their sliders. This ensures that all sensors have a unified, stable, and high-precision circumferential measurement reference, avoiding the cumulative errors and relative vibrations that may occur when sensors are mounted on multiple independent cantilever arms or supports.

[0032] Specifically, the lower half-ring fixed base 81 sits on two high-precision linear guide rails 14 that are parallel to each other on the base 1 via four sliders at its bottom, and can be driven by a servo motor to move horizontally along the X-axis; the upper half-ring movable cover 82 is connected to the lower half-ring via a heavy-duty hinge and is closed and locked by an electromagnetic lock.

[0033] In some embodiments, such as Figure 2 As shown, the drive assembly 7 is a pneumatic turbine drive ring fixed to the base 1 and located below the turbine impeller 5. It includes multiple miniature high-speed solenoid valve nozzles 12 that are evenly distributed circumferentially and have adjustable injection directions. Specifically, there are 24 miniature high-speed solenoid valve nozzles 12, each of which is mounted via a graduated ball joint, allowing adjustment of its jet airflow direction to achieve an optimal tangential angle (approximately 20°) with the impeller blades. All nozzles are controlled by the control system 2 using PWM.

[0034] In this embodiment, a pneumatic turbine drive ring composed of circumferentially distributed miniature high-speed solenoid valve nozzles 12 is used. This ring directly acts on the impeller blades by injecting high-speed airflow, providing rotational torque. This drive method achieves complete physical isolation (non-contact) between the drive source and the impeller's rotating shaft system, eliminating all additional vibrations and noise from the mechanical transmission chain. This creates ideal conditions for acoustic, vibration, and high-precision displacement sensors 48 to collect the impeller's "pure" intrinsic dynamic response signal. Simultaneously, the high-speed solenoid valve nozzles feature an adjustable injection direction design. After changing the impeller specifications, the angle of the nozzle airflow relative to the blades can be changed manually or automatically to find the optimal drive efficiency point. Combined with the control system 2's adjustment of airflow pressure and pulse frequency, the optimal drive parameters can be quickly matched for different impellers, ensuring stable and efficient rotation. Furthermore, each miniature high-speed solenoid valve nozzle 12 can be independently controlled by the control system 2 using high-frequency pulse width modulation (PWM). By adjusting the nozzle's opening time, frequency, and combination sequence, the total driving torque acting on the impeller can be controlled with great flexibility, thereby achieving a wide range of stepless smooth speed regulation from extremely low speeds to high speeds. In addition, by individually adjusting the injection parameters of nozzles at specific angles, the unbalanced torque of the impeller can be compensated to a certain extent, assisting in achieving smoother start-up, shutdown, and rotation.

[0035] In some embodiments, such as Figure 6As shown, it also includes a non-contact braking assembly 13, which includes a brake disc 131 coaxially connected to the main shaft and a brake 132 fixed on the base 1. Specifically, the brake 132 is a C-type iron-core electromagnetic brake 132, which maintains a gap of about 1 mm with the edge of the brake disc 131 to achieve non-contact eddy current braking.

[0036] This embodiment employs non-contact braking (such as eddy current braking or hysteresis braking). The braking force originates from the interaction between the electromagnetic field and the moving conductor (i.e., brake disc 131). A physical gap is maintained between the brake 132 and the rotating brake disc 131, eliminating any contact friction and preventing wear and debris generation. This ensures the high cleanliness of the detection chamber and guarantees the long-term stable operation of the air bearing and optical sensor. Furthermore, the braking torque of non-contact braking (especially eddy current braking) is proportional to the rotational speed within a certain range, achieving a natural and smooth braking characteristic that automatically decreases as the rotational speed decreases. This avoids the impact load and rapid deceleration that may result from mechanical braking. This smooth braking process protects the precision impeller and main shaft support system from impact damage. Moreover, through coordination with the control system 2, precise programming of the braking curve (such as constant deceleration braking) can be achieved, thereby accurately controlling the stopping time and position. Combined with non-contact drive, the entire device's "start-stabilize-brake" cycle can be completed quickly and smoothly, significantly shortening the auxiliary detection time for a single product and improving overall efficiency.

[0037] Specifically, the double-hemispherical air bearing achieves contactless support, the pneumatic turbine drive ring achieves contactless actuation, and the non-contact braking component 13 perfectly closes the contactless braking loop. These three components together constitute a logically self-consistent and technologically unified fully contactless working chain. This not only avoids compatibility issues that may arise from different types of interactions (contact / non-contact) in engineering, but also ensures, conceptually, that the impeller remains in a near-ideal state free from additional mechanical interference throughout the entire process from startup, operation to shutdown.

[0038] In some embodiments, such as Figure 7 As shown, the quick-change fixture 6 is a hydraulic expansion mandrel, which is externally fitted with replaceable modular adapters 15. Specifically, the hydraulic expansion mandrel of the quick-change fixture 6 is connected to the front end of the main shaft inside the air bearing via a high-rigidity diaphragm coupling. A series of modular adapters 15 are provided for turbine impellers 5 with different inner diameters. During clamping, the appropriate adapter is fitted onto the mandrel, the impeller is fitted over the adapter, and the hydraulic system is activated to expand the mandrel. The adapter then evenly clamps the impeller through the inner hole, achieving non-destructive, high concentricity clamping.

[0039] This embodiment uses a hydraulic expansion mandrel as the core clamping mechanism. When the hydraulic system is pressurized, the mandrel generates uniform radial micro-expansion, which, through the replaceable modular adapter 15, clamps the impeller's inner bore in a 360-degree, gapless manner. This uniform expansion force, without eccentric loading, avoids the impeller's inner bore deformation, slight eccentricity, or surface indentation that can occur with traditional three-jaw chucks due to single-point or three-point force application. For impellers with different inner diameter specifications, only a corresponding lightweight adapter needs to be replaced. During operation, the hydraulic pressure is released, the old adapter is removed, the new adapter is installed, and pressure is reapplied. The entire process requires no tools to adjust the jaw position or perform complex centering calibration and can usually be completed within 1-3 minutes.

[0040] In some embodiments, such as Figure 4 , Figure 5 As shown, the intelligent sensor group 9 includes at least two of the following: a laser displacement sensor 91, a 3D scanner 92, and an industrial camera 93. The laser displacement sensor 91 and the 3D scanner 92 are mounted on the arc-shaped guide rail 85 via sliders 84 at their bottoms and are fixed in position by locking knobs on the sides. When the aviation plug (i.e., electrical plug 11) on the side of the module is pushed into place, it automatically connects to the corresponding socket (i.e., standard electrical interface 86) inside the ring wall.

[0041] The industrial camera 93 is mounted on a slide table 94 that can move along the arc-shaped guide rail 85. The slide table 94 integrates a servo drive mechanism (including a micro servo motor and a ball screw), which can perform continuous circumferential or point-to-point movement on the arc-shaped guide rail 85. The air-bearing vibration isolation unit 3 has a honeycomb sandwich structure with interconnected honeycomb-shaped air chambers inside. Specifically, the upper and lower parts are sealed steel plates, and the middle part is an aluminum hexagonal honeycomb core with an inscribed circle diameter of approximately 30 mm. After an external low-pressure air source (approximately 0.05 MPa) is introduced, the gas is evenly distributed in the interconnected honeycomb-shaped air chambers, making the entire base 11 suspend on the ground and effectively isolating environmental vibrations.

[0042] Specifically, the lower half-ring fixing base 81 is provided with two parallel arc-shaped guide rails 85, one of which is equipped with a laser displacement sensor 91 and a 3D scanner 92 for manual adjustment, and the other is equipped with an industrial camera 93 and a servo drive mechanism.

[0043] In this embodiment, the laser displacement sensor 91 and the 3D scanner 92 are responsible for performing a highly efficient and accurate "comprehensive survey" of the macroscopic contour, radial runout, and three-dimensional morphology of the entire blade group when the impeller rotates, quickly locating suspicious areas. Based on this, the slide table 94, with its integrated servo drive mechanism, can move precisely along the arc-shaped guide rail 85, quickly and accurately driving the high-resolution industrial camera 93 to the angular position of the suspected defect point for fixed-point high-definition imaging or video recording, achieving "fixed-point precision inspection." This working mode balances detection efficiency with the accuracy and detail richness of defect identification, and is particularly suitable for the precise capture and classification of complex and minute surface defects (such as microcracks and ablation points).

[0044] Furthermore, mounting the industrial camera 93 on a servo slide 94 capable of precise circumferential movement endows the vision system with unprecedented flexibility. It can perform automated point-to-point inspection of key parts of each blade according to a preset program, or adaptively perform focused re-inspection of specific areas based on feedback from other sensors (such as the 3D scanner 92). This dynamic scanning capability ensures comprehensive visual coverage of the impeller surface, significantly improving the defect detection rate.

[0045] Furthermore, the honeycomb sandwich structure air-bearing vibration isolation unit 3, with its interconnected honeycomb air chambers, forms an "air spring" matrix with excellent damping characteristics, isotropic stiffness, and uniform load-bearing capacity after low-pressure air is introduced. Compared with simple air cushions or rubber vibration isolators, the honeycomb structure can more effectively attenuate and isolate broadband vibrations (especially low-frequency vibrations) transmitted from the ground, providing a near-"static" reference platform for the entire detection device and eliminating the adverse effects of environmental vibrations on the reading stability of the high-precision displacement sensor 48, the clarity of the 3D scanning point cloud, and the quality of long-exposure images.

[0046] In some embodiments, each sensor is used to collect multi-source signals during the rotation of the turbine impeller 5. The control system 2 fuses the collected multi-source spatiotemporal synchronous data and automatically identifies defects and generates a comprehensive inspection report through a pre-trained AI model. The multi-source signals include vibration signals, three-dimensional point cloud data, surface image data, infrared thermal imaging data, and acoustic emission signals. The AI ​​model is a deep learning model based on multi-source signal fusion. The infrared thermal imaging data is collected by an infrared imager, which is installed at the top center of the inner side of the upper semi-ring movable cover 82, with the lens vertically downward or slightly tilted towards the impeller.

[0047] In this embodiment, a pre-trained multi-source signal fusion deep learning AI model is used to perform end-to-end automated analysis of synchronously acquired vibration, 3D morphology, images, thermal images, and acoustic emission signals. This model can autonomously discover the implicit and complex correlations between different physical signals. For example, it can correlate vibration anomalies at specific frequencies with minute contour deviations, hot spots, and specific acoustic emission events in a local area of ​​a blade. This enables highly accurate and reliable automatic identification and classification of defects (such as cracks, material defects, and assembly anomalies). Furthermore, through multi-source spatiotemporal synchronous data fusion, it has for the first time constructed a complete digital profile reflecting the coupling state of multiple physical fields (mechanical, thermal, acoustic, and geometric) of the impeller under simulated operating conditions. The AI ​​model can not only identify existing macroscopic defects but also sensitively capture early signs of potential failures such as material performance degradation, microcrack initiation, and coating peeling by analyzing subtle synergistic changes between multiple signals (such as increased acoustic emission activity in a specific location accompanied by minor local thermal distribution anomalies). This achieves predictive health assessment and early warning capabilities far exceeding traditional methods.

[0048] Specifically, control system 2 integrates an industrial PC, PLC, high-speed data acquisition card, proportional valve controller, etc. It is responsible for coordinating all motion control (detection loop opening and closing movement, slide table (94 motion, pneumatic drive / braking), synchronous data acquisition (synchronously triggering all sensors according to the spindle encoder signal), and running an AI defect recognition model based on deep learning.

[0049] Working principle and testing process: S1: The control system 2 controls the electromagnetic lock to open, and opens the upper half ring movable cover 82 through the hinge, so that the detection ring forms a "C" shaped structure of about 270°, and then drives the entire segmented detection ring 8 to move horizontally out of the central area along the linear guide rail.

[0050] S2: The operator vertically hoists the turbine impeller 5 to be tested onto the fitting of the quick-change fixture 6 and starts the hydraulic locking.

[0051] S3: Drive the detection ring back to the center, realigning the open "C" shaped notch with the impeller. Then drive the lower half ring to move, allowing the impeller to fully enter the space enclosed by the inner side of the lower half ring and the open upper half ring. Close the upper half ring and lock it. At this point, the impeller is located at the center of the closed detection ring.

[0052] S4: Start the air supply system to supply air (pressure 0.3-0.6 MPa) to the double hemisphere air bearing assembly 4, suspending the impeller main shaft. The control system 2 reads the gap values ​​of the four displacement sensors 48 and dynamically adjusts the output pressure of the four proportional valves through a PID algorithm to keep the gaps in the four regions consistent, thus completing active leveling.

[0053] S5: Start the pneumatic turbine drive ring and control the nozzle according to the preset program to drive the impeller to the target speed (such as the rated operating speed). During rotation, the control system 2 synchronously triggers the laser displacement sensor 91 (to collect radial vibration for dynamic balance analysis), the 3D scanner 92 (to acquire the three-dimensional morphology of the blade), the industrial camera 93 (to capture surface images by fixing points or scanning), the infrared thermal imager (to monitor the temperature field), and the embedded acoustic emission sensor (to capture internal damage signals).

[0054] S6: After the test is completed, activate the eddy current brake 132 to bring the turbine impeller 5 to a smooth and rapid stop.

[0055] S7: Control system 2 inputs all collected multi-source spatiotemporal synchronous data into a pre-trained multi-source information fusion deep learning model (e.g., an improved convolutional neural network with 3D point clouds, vibration spectra, acoustic emission waveform features, and thermal images as inputs). The model automatically outputs defect identification results (including type, location, and size rating) and automatically generates an interactive digital twin inspection report containing 3D defect annotations.

[0056] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0057] Furthermore, the use of terms such as "first" and "second" in this application is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed in this application.

[0058] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0059] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A device for detecting and identifying air bearing turbine impellers, characterized in that, The system includes a base (1) and a control system (2). The base (1) has an air-bearing vibration isolation unit (3) at its bottom and a double hemisphere air bearing assembly (4) for suspending and supporting the main shaft of the turbine impeller (5) in the central area of ​​its upper surface. The front end of the main shaft of the double hemisphere air bearing assembly (4) is connected to a quick-change clamp (6) for clamping the turbine impeller (5). A non-contact drive assembly (7) for driving the suspended turbine impeller (5) to rotate is installed on one side of the double hemisphere air bearing assembly (4). A split detection ring (9) is slidably arranged on the base (1). The split detection ring (9) can move horizontally to cover or avoid the turbine impeller (5). The split detection ring (9) is equipped with a smart sensor group (9) for detection. An acoustic emission sensor group (10) is embedded inside the double hemisphere air bearing assembly (4). The smart sensor group (9), the acoustic emission sensor group (10) and the drive assembly (7) are all electrically connected to the control system (2).

2. The air bearing turbine impeller detection and identification device as described in claim 1, characterized in that, The double hemisphere air bearing assembly (4) includes a bearing housing (41) and an upper hemisphere bearing (42) and a lower hemisphere bearing (43) installed in the bearing housing (41); the bearing housing (41) is engaged with the positioning stop on the base (1), and at least two annular grooves (44) are machined on the inner spherical surface of the upper hemisphere bearing (42) and the lower hemisphere bearing (43), which divide the inner surface of each hemisphere bearing into at least three independent pressure zones (45); each pressure zone (45) is provided with an independent air intake channel (46) on its back; each air intake channel (46) is externally connected to an electric proportional valve (47) controlled by the control system (2), and each pressure zone (45) is provided with a displacement sensor (48) for monitoring the impeller journal suspension clearance.

3. The air bearing turbine impeller detection and identification device as described in claim 2, characterized in that, The acoustic emission sensor group (10) is embedded in the side wall of the bearing housing (41) with an interference fit, and its sensing end face is flush with the inner cavity of the bearing.

4. The air bearing turbine impeller detection and identification device as described in claim 3, characterized in that, The split detection ring (9) includes a lower half-ring fixing base (91), an upper half-ring movable cover (92) hinged thereto, and a locking mechanism (93) that fixes the two together; the lower half-ring fixing base (91) is provided with an arc-shaped guide rail (95) along the circumferential direction and a standard electrical interface (96) on its inner side; each sensor of the intelligent sensor group (9) can be detachably installed on the arc-shaped guide rail (95) by the slider (94) at its bottom, and its position is fixed by the locking knob; when it is in place during installation, the electrical plug (11) on the sensor is automatically plugged into the standard electrical interface (96).

5. The air bearing turbine impeller detection and identification device as described in claim 4, characterized in that, The drive assembly (7) is a pneumatic turbine drive ring fixed to the base (1) and located below the turbine impeller (5), which includes a plurality of miniature high-speed solenoid valve nozzles (12) that are evenly distributed in the circumference and whose injection direction is adjustable.

6. The air bearing turbine impeller detection and identification device as described in claim 1, characterized in that, It also includes a non-contact braking assembly (13), which includes a brake disc (131) coaxially connected to the main shaft and a brake (132) fixed to the base (1).

7. The air bearing turbine impeller detection and identification device as described in claim 6, characterized in that, The quick-change clamp (6) is a hydraulic expansion mandrel, which is fitted with a replaceable modular adapter (15).

8. The air bearing turbine impeller detection and identification device as described in claim 7, characterized in that, The intelligent sensor group (9) includes at least two of the following: a laser displacement sensor (91), a 3D scanner (92), and an industrial camera (93).

9. The air bearing turbine impeller detection and identification device as described in claim 8, characterized in that, The industrial camera (93) is mounted on a slide (94) that can move along the arc-shaped guide rail (95). The slide (94) is integrated with a servo drive mechanism and can make continuous or point-to-point circumferential movements on the arc-shaped guide rail (95). The air-bearing vibration isolation unit (3) is a honeycomb sandwich structure with interconnected honeycomb air chambers inside.

10. The air bearing turbine impeller detection and identification device as described in claim 9, characterized in that, Each sensor is used to collect multi-source signals during the rotation of the turbine impeller (5). The control system (2) fuses the collected multi-source spatiotemporal synchronous data and automatically identifies defects and generates a comprehensive inspection report through a pre-trained AI model. The multi-source signals include vibration signals, three-dimensional point cloud data, surface image data, infrared thermal image data and acoustic emission signals. The AI ​​model is a deep learning model based on the fusion of multi-source signals.

Images