Aero-engine defective blade detection device and method based on laser repair

Abstract

The invention discloses an aero-engine defective blade detection device and method based on laser repair, and the device comprises a blue light three-dimensional scanner, a two-dimensional image collection system, and a blade three-dimensional calibration rotating clamp, and the blue light three-dimensional scanner and the two-dimensional image collection system are disposed at the periphery of the blade three-dimensional calibration rotating clamp. The blade three-dimensional calibration rotary clamp is used for fixing a detection target, and aiming at laser repair of large-scale defects on the surface of the aero-engine blade, synchronous data acquisition of a two-dimensional image and a three-dimensional model can be realized, and the detection accuracy and the detection efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the field of machine vision defect detection, and in particular relates to a device and method for detecting defective blades of an aero-engine based on laser repair. Background Art

[0002] The "one disc and two blades" are the most core components of an aircraft engine. Specifically, they refer to the turbine disc, guide vanes, and rotor blades. The rotor blades primarily include compressor blades and turbine blades. Blade parts are considered one of the most critical components in routine aircraft engine maintenance and repair.

[0003] When working, they have to withstand high temperatures, alternating mechanical loads with large amplitude and frequency changes, thermal alternation and vibration. The working environment is extremely harsh. Under the influence of centrifugal force, thermal stress and mechanical stress generated by temperature gradient, the blades will experience creep, fatigue, thermal corrosion and other failure forms; due to the large number and thin blade shape, the impact of external particles and flying birds can easily cause the blades to form pits, material loss, scratches, corrosion, torsion and other large-scale defects, which reduce the overall performance and reliability of the engine and seriously affect its service life.

[0004] Laser precision repair technology, a key technology for repairing damaged aero-engine blades, demonstrates excellent repair results. This technology is a relatively advanced laser surface repair / strengthening technique developed in recent years. Through synchronized powder feeding, a high-energy beam is used above the molten pool to simultaneously melt the additive material and the high-speed moving surface of the base material. This material then rapidly solidifies to form a cladding layer with an extremely low dilution rate and a metallurgical bond to the base. Compared to traditional repair techniques such as manual argon arc welding, arc additive manufacturing, and plasma spraying, this technology boasts high efficiency, low surface roughness of the cladding layer, a small heat-affected zone, and minimal base deformation. This technology addresses a series of technical challenges that are difficult to overcome with traditional surface repair techniques, including material selection limitations, large base thermal deformation, coarse microstructure, poor thermal fatigue resistance, and poor interface bonding strength. Therefore, it is more suitable for the repair and reshaping of complex, high-temperature, thin-walled metal aero-engine blades.

[0005] However, before repairing an aeroengine, it is necessary to accurately detect the repair defects and accurately extract the defect characteristic parameters to guide laser cladding path planning and trajectory optimization. Traditional methods or the use of machine vision target detection alone are unable to accurately detect the types of defects in aeroengine blades, especially those caused by twisted blades. This results in inefficient and costly laser repair of defective aeroengine blades. Summary of the Invention

[0006] To solve the above problems, the present invention provides an aircraft engine defective blade detection device based on laser repair. The device has the advantages of high defect recognition accuracy, accurate positioning, and precise and fast grading of defective blades. It is of great significance to reduce the number of parts replacements, reduce costs, shorten production cycles, and extend the service life of equipment in the aviation field.

[0007] Another object of the present invention is to provide a detection method suitable for the above device.

[0008] The technical solution adopted by the present invention is a device for detecting defective blades of an aero-engine based on laser repair, comprising a blue light 3D scanner, a blade 3D calibration rotating fixture, a 2D image acquisition system, a computer, The blade three-dimensional calibration rotation fixture includes a rotating table, a calibration plate, a fixture table, and a circular guide ring. The top surface of the rotating table is fixedly connected to the calibration plate, and the top surface of the calibration plate is fixedly connected to the fixture table. The rotating table is arranged on the inner side of the circular guide ring. The rotating table, calibration plate, fixture table, and circular guide ring are concentrically arranged. The rotating table circuit is connected to a control panel. The two-dimensional image acquisition system includes a macro lens industrial camera and two sets of LED strip light source systems. The macro lens industrial camera is installed on one side of the three-dimensional calibration rotating fixture and connected to the computer circuit. The two sets of LED strip light sources are respectively installed on circular guide rails and placed at 75° on both sides of the industrial camera. The circuits are connected to the light source controller. The blue light 3D scanner is mounted on a tripod and placed on the other side of the blade 3D calibration rotating fixture and connected to the computer. The computer is provided with an aero-engine blade defect detection system.

[0009] The blue light 3D scanner has binocular cameras on both sides of the front end, a grating sensor in the middle, a near-end field of view of 145×120-340mm, a measuring point distance of 0.11mm, a single-frame measurement accuracy of 0.02mm, a single scanning time of 0.6s, and a device size of 280×200×120mm.

[0010] The industrial camera is a CCD global exposure color camera with a macro lens, a camera resolution of 2448×2048 pixels, a Sony 2 / 3" CMOS sensor, a chip size of 8.8×6.6mm, a maximum frame rate of 20fps, and a global exposure mode. The field of view is 36.2×27.2 mm.

[0011] The LED strip light source is a white light source with a size of 240×30 mm and is arranged vertically.

[0012] The device of the present invention is low-cost, easy to reproduce, small in size and easy to carry, with a high degree of system modularity. Each module uses a universal interface, which is convenient for transplantation. It uses physical drive, has a fast detection time, intuitive data display, and can record information throughout the process. It can be used for laser repair of aircraft engine blade assembly line defects and is easy to integrate online.

[0013] A method for detecting defective blades of an aero-engine based on laser repair comprises the following steps: S1. Adjust the position of the LED strip light source and select different light source intensities according to the material and size of the test target, and fix the test target on the fixture table of the blade 3D calibration rotating fixture; S2. Turn on the industrial camera and the rotating stage, and observe and adjust them on the computer display interface to obtain a clear image of the inspection target; S3, use computer to process S2 to obtain image data, extract image feature values, and calculate the accuracy and overlap (IOU) of the detected target. Accuracy , Where TP is the number of positive class predictions, FP is the number of negative class predictions, and N is the total number of predictions. Overlap IOU = inter_area / union_area, Where inter_area represents the overlapping area of ​​the two prediction boxes in machine vision inter_area = (yi2 -yi1) * (xi2 - xi1), and union_area represents the union area of ​​the two prediction boxes in machine vision union_area = box1_area + box2_area - inter_area; S4. According to the accuracy and IOU, obtain and record the defect type, quantity, and location information of the detection target. S5. Reconstruct a three-dimensional model of the inspection target, compare it with the intact target model, calculate the offset of the reconstructed model, and classify the defects of the aircraft engine blade.

[0014] The specific calculation method of the offset of S5 is: 1) Turn on the blue light 3D scanner for 3D calibration; 2) Turn on the rotating stage, set the rotation speed to 3 rad / min, and the rotation time to 20 s; 3) Use a blue light scanner to scan the inspection target with large-scale surface defects to obtain the three-dimensional point cloud data of the blade; 4) The computer processes the 3D point cloud data obtained in step 3 and compares it with the existing intact 3D point cloud data of the target model. The surface normal vector and curvature are estimated based on the neighborhood points. Then, the superimposable points between the measured point cloud and the target point cloud are matched based on the curvature. All normal vector directions in the superimposable point set are mapped into a consistent 3D space transformation. The rotation and translation transformation rules of the Cartesian coordinate system are used to calculate the rotation matrix corresponding to the rotation of the coordinates around the X, Y, and Z axes. , Then the rotation matrix can be expressed as ,

[0015] In the formula Respectively represent the rotation angles of the coordinate point around the X, Y, and Z axes. The torsion offset can be expressed as shown in the formula: .

[0016] The detection method of the present invention is based on the two-dimensional aircraft engine blade defect detection of traditional machine vision, based on laser repair technology and combined with three-dimensional reconstruction. It can fully understand the shape, size, and morphological structure of the target defect, and realize accurate identification, positioning, and classification of aircraft engine blade defects. At the same time, the two-dimensional image and three-dimensional model of the blade defect are extracted to obtain a transplantable, portable, and callable database, which provides a model basis for the next step of laser repair path planning and numerical simulation, and provides a data basis for evaluating the performance of aircraft engine blades in different working environments and further optimizing their manufacturing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the structure and working of the patent of this invention In the figure: 1. Blue light 3D scanner; 2. Macro lens industrial camera; 3. LED strip light source; 4. Light source controller; 5. Tripod; 6. Computer; 7. Circular guide rail; 8. Fixture table; 9. Calibration plate; 10. Rotation table; 11. Control panel. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, a device for detecting defective blades of an aircraft engine based on laser repair includes a blue light 3D scanner 1, a blade 3D calibration rotating fixture, a 2D image acquisition system, a computer 6, The blade three-dimensional calibration rotation fixture includes a rotating table 10, a calibration plate 9, a fixture table 8, and a circular guide ring 7. The top surface of the rotating table 10 is fixedly connected to the calibration plate 9, and the top surface of the calibration plate 9 is fixedly connected to the fixture table 8. The rotating table 10 is arranged on the inner side of the circular guide ring 7. The rotating table 10, calibration plate 9, fixture table 8, and circular guide ring 7 are concentrically arranged. The rotating table 10 is connected to a control panel 11. The two-dimensional image acquisition system includes a macro lens industrial camera 2 and two sets of LED bar light systems 3. The industrial camera is a CCD global exposure color camera with a macro lens. The camera has a resolution of 2448 × 2048 pixels and is equipped with a Sony 2 / 3" CMOS sensor with a chip size of 8.8 × 6.6 mm. It has a maximum frame rate of 20 fps and uses global exposure mode. The field of view range is 36.2 × 27.2 mm. The LED bar light system is a white light source with a size of 240 × 30 mm. The macro lens industrial camera 2 is installed on one side of the three-dimensional calibration rotating fixture and connected to the computer 6. Two sets of LED strip light sources 3 are longitudinally installed on the circular guide rail 7, respectively placed at 75 degrees on both sides of the industrial camera, and connected to the light source controller 4. The blue-light 3D scanner 1 is mounted on a tripod 5 and placed on the other side of the blade 3D calibration rotating fixture and connected to a computer 6. The blue-light 3D scanner has binocular cameras on both sides of the front face and a grating sensor in the middle. The near-end field of view is 145×120-340mm, the measurement point spacing is 0.11mm, the single-frame measurement accuracy is 0.02mm, the single scan time is 0.6s, and the device dimensions are 280×200×120mm. The computer 6 is provided with an aircraft engine blade defect detection system.

[0020] In order to obtain the category, level and location information of defects from the two-dimensional image and three-dimensional point cloud model data of defective aero-engine blades, the aero-engine defective blade detection method based on laser repair is as follows:

[0021] First, turn on the LED strip light source 3 and set the illumination of the light source to 200 Klux; place the aircraft engine blade on the table fixture 8; Connect the macro lens industrial camera 2, turn on the rotating stage 10 through the rotating stage control panel 11, observe and adjust on the display interface of the computer 6 to obtain a clear image; The image data is processed by computer 6, the image feature values ​​are extracted, and the accuracy and overlap of the detection target are calculated by combining with the aircraft engine blade defect detection system. The accuracy calculation formula is: , Where TP is the number of positive classes predicted as positive, FP is the number of negative classes predicted as positive, and N is the total number of predictions; The calculation formula for coincidence degree is: IOU= inter_area / union_area, Where inter_area represents the overlapping area of ​​the two prediction boxes in machine vision inter_area = (yi2 -yi1) * (xi2 - xi1), and union_area represents the union area of ​​the two prediction boxes in machine vision union_area = box1_area + box2_area - inter_area; Then, the defect detection system parameters are adjusted according to the accuracy and IOU size, and the defect type, quantity, and location information of the defective blades of the aircraft engine are recorded.

[0022] Reconstruct the three-dimensional model of the aircraft engine blade and calculate the torsion offset of the aircraft engine blade. Turn on the blue light 3D scanner 1 for 3D calibration; Use the rotating stage operation panel to turn on the rotating stage, set the rotation speed to 3 rad / min, and the rotation time to 20 s; Use a blue light scanner to scan an aircraft engine blade with large surface defects to obtain three-dimensional point cloud data of the blade. The computer processes the blade 3D point cloud data. Based on the existing complete blade 3D point cloud data, the surface normal vector and curvature are estimated according to the neighborhood points. Then, the superimposable points between the measured point cloud and the target point cloud are matched according to the curvature. All normal vector directions in the superimposable point set are mapped into a consistent 3D space transformation. The rotation and translation transformation rules of the Cartesian coordinate system are used to calculate the rotation matrix corresponding to the coordinate rotation around the X, Y, and Z axes. , Then the rotation matrix can be expressed as , In the formula Respectively represent the rotation angles of the coordinate point around the X, Y, and Z axes. The translation matrix, i.e. the torsion offset, can be expressed as follows: , According to the offset of the defective blade model, the defect grade of the defective aero-engine blade is classified based on the laser repair process of aero-engine blades.

[0023] At this point, the accurate defect type, level, and location of the aircraft engine blades are calculated through two-dimensional images and three-dimensional point cloud data. In summary, the aircraft engine defect blade detection device and method based on laser repair of the present invention, through the acquisition of two-dimensional images of aircraft engine blades, according to the image feature information, for aircraft engine blades prone to form large-scale surface defects such as pits, material loss, scratches, corrosion, wear and tear, high-precision identification; using blue light three-dimensional scanning technology, obtain the three-dimensional point cloud data of the aircraft engine defective blade, by comparing with the complete blade three-dimensional point cloud data, calculate the offset T of the point cloud data in the XYZ axis, and use the point cloud registration function to calculate The blade torsional offset is determined and the blade defect grade is classified. The purpose of the classification is to exclude aircraft engine blades with large torsional offsets, that is, deformation failures that cannot be repaired, and select aircraft engine blades with deformation within the specified range that can be repaired and continued to be used after laser repair.

[0024] The above only describes in detail the preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Various changes can be made within the knowledge of ordinary technicians in this field without departing from the purpose of the present invention, and various changes should be included in the scope of protection of the present invention.

Claims

1. A device for detecting defective blades of an aero-engine based on laser repair, characterized in that: Including blue light 3D scanner, blade 3D calibration rotation fixture, 2D image acquisition system, computer, The blade three-dimensional calibration rotation fixture includes a rotating table, a calibration plate, a fixture table, and a circular guide ring. The top surface of the rotating table is fixedly connected to the calibration plate, and the top surface of the calibration plate is fixedly connected to the fixture table. The rotating table is arranged on the inner side of the circular guide ring. The rotating table, calibration plate, fixture table, and circular guide ring are concentrically arranged. The rotating table circuit is connected to a control panel. The two-dimensional image acquisition system includes a macro lens industrial camera and two sets of LED strip light source systems. The macro lens industrial camera is installed on one side of the three-dimensional calibration rotating fixture and connected to the computer circuit. The two sets of LED strip light sources are respectively installed on circular guide rails and placed at 75° on both sides of the industrial camera. The circuits are connected to the light source controller. The blue light 3D scanner is mounted on a tripod and placed on the other side of the blade 3D calibration rotating fixture and connected to the computer. The computer is provided with an aero-engine blade defect detection system.

2. The device for detecting defective blades of an aero-engine based on laser repair according to claim 1, characterized in that: The blue light 3D scanner has binocular cameras on both sides of the front end, a grating sensor in the middle, a near-end field of view of 145×120-340mm, a measuring point distance of 0.11mm, a single-frame measurement accuracy of 0.02mm, and a single scanning time of 0.6s.

3. The device for detecting defective blades of an aero-engine based on laser repair according to claim 1, characterized in that: The industrial camera is a CCD global exposure color camera with a macro lens, a camera resolution of 2448×2048 pixels, a Sony 2 / 3" CMOS sensor, a maximum frame rate of 20fps, and a field of view of 36.2×27.2 mm.

4. The device for detecting defective blades of an aero-engine based on laser repair according to claim 1, characterized in that: The LED strip light source is a white light source and is arranged vertically.

5. The device for detecting defective blades of an aero-engine based on laser repair according to claim 1, characterized in that: The detection method of the device comprises the following steps: S1. Adjust the position of the LED strip light source and select different light source intensities according to the material and size of the test target, and fix the test target on the fixture table of the blade 3D calibration rotating fixture; S2. Turn on the industrial camera and the rotating stage, and observe and adjust them on the computer display interface to obtain a clear image of the inspection target; S3, use computer to process S2 to obtain image data, extract image feature values, and calculate the accuracy and overlap (IOU) of the detected target. Accuracy , Where TP is the number of positive class predictions, FP is the number of negative class predictions, and N is the total number of predictions. Overlap IOU = inter_area / union_area, Where inter_area represents the overlapping area of ​​the two prediction boxes in machine vision inter_area = (yi2 - yi1) * (xi2 - xi1), and union_area represents the union area of ​​the two prediction boxes in machine vision union_area = box1_area + box2_area - inter_area; S4. According to the accuracy and IOU, obtain and record the defect type, quantity, and location information of the detection target. S5. Reconstruct a three-dimensional model of the detected target, compare it with the intact target model, calculate the offset of the reconstructed model, and classify the target defects.

6. The device for detecting defective blades of an aero-engine based on laser repair according to claim 5, characterized in that: The offset calculation method of S5 is: 1) Turn on the blue light 3D scanner for 3D calibration; 2) Turn on the rotating stage, set the rotation speed to 3 rad / min, and the rotation time to 20 s; 3) Use a blue light scanner to scan the target with large-scale surface defects to obtain the target's three-dimensional point cloud data; 4) The computer processes the 3D point cloud data obtained in step 3 and compares it with the 3D point cloud data of the existing intact target model. The surface normal vector and curvature are estimated based on the neighborhood points. Then, the superimposable points between the measured point cloud and the target point cloud are matched based on the curvature. All normal vector directions in the superimposable point set are mapped into a consistent 3D space transformation. The rotation and translation transformation rules of the Cartesian coordinate system are used to calculate the rotation matrix corresponding to the rotation of the coordinates around the X, Y, and Z axes. ， Then the rotation matrix can be expressed as ， In the formula Respectively represent the rotation angles of the coordinate point around the X, Y, and Z axes. The torsion offset can be expressed as shown in the formula: 。

Images