A method and system for detecting deformation of engine vanes

Abstract

The application discloses a kind of method and system for detecting engine blade deformation, belong to engine detection field.For the problem of slow engine blade detection efficiency and poor accuracy, the application provides a kind of method for detecting engine blade deformation, comprising the following steps: obtaining original image, including the ball cap part in engine and the blade part to be detected in original image;Respectively extract the outline of ball cap and the blade outline to be detected in original image;According to the ideal blade outline of the extracted ball cap outline transformation, calculate the blade to be detected;Compare ideal blade outline with the extracted blade outline to be detected, judge whether deformation occurs.The application compares ideal blade outline with the blade outline to be detected to determine whether the blade is deformed, the accuracy is effectively guaranteed, the degree of automation is high, and the detection efficiency is improved.The system structure of the application is simple, which can realize efficient inspection of blade deformation and high-precision detection of blade deformation.

Description

Technical Field

[0001] This invention belongs to the field of engine testing, and more specifically, relates to a method and system for detecting engine blade deformation. Background Technology

[0002] Aero engines are the core power source of aircraft, and their complex and precise structure showcases the high standards of intelligent aircraft manufacturing technology. Engine blades, as crucial components for engine performance, often have complex shapes and operate under harsh conditions year-round, subject to centrifugal loads, aerodynamic loads, temperature loads, impact loads, and environmental corrosion. Blades are one of the core components of aero engines, and their quality inspection and life prediction are fundamental to the safe operation of the engine. Currently, most engine blade inspections rely on visual inspection, which suffers from inaccuracies, lack of quantifiable indicators, and potential oversights leading to safety hazards. Alternatively, laser sensors can be used for blade measurement; while laser positioning is accurate, it is expensive and not ideal for extracting minute features on the blade.

[0003] To address the aforementioned issues, corresponding improvements have been made. For example, Chinese patent application CN202111051068.9, published on December 3, 2021, discloses an integrated detection method for cracks and residual stress in aero-engine blades. Based on infrared thermal imaging and binocular vision technology, it enables rapid in-situ detection of cracks and residual stress in blades without disassembling the aero-engine. The method uses a computer-controlled array of hot air nozzles or a single hot air nozzle to generate hot airflow to thermally excite the aero-engine blades. Under this excitation, the release of different residual stresses leads to varying degrees of torsional and bending deformations in the blades. Simultaneously, cracks cause abnormal temperature responses on the blade surface. A binocular vision camera and an infrared thermal imager are used to detect the thermal deformation and temperature response of the blades during thermal excitation. The magnitude of residual stress and the extent of cracking are evaluated based on the degree of deformation and temperature response characteristics. The drawbacks of this patent are its slow detection efficiency and high cost.

[0004] For example, Chinese patent application CN202010495253.6, published on October 13, 2020, discloses a method for detecting aero-engine blades based on an RGB-D camera. This method involves scanning engine blades with an RGB-D camera to acquire RGB-D data, removing noise from the depth image using bilateral Bayesian filtering, estimating the extrinsic parameter T of the RGB-D sensor using the depth data, calculating the RGB-D point cloud based on the sensor's intrinsic parameter K, and then stitching the point cloud together using the sensor's extrinsic parameter T to form a preliminary blade point cloud image. Finally, downsampling filtering and extrinsic point removal filtering are used to optimize the preliminary blade point cloud image, yielding the final blade point cloud image. The drawback of this patent is that the detection accuracy cannot be effectively guaranteed. Summary of the Invention

[0005] 1. The problem to be solved

[0006] To address the problems of slow efficiency and poor accuracy in engine blade inspection, this invention provides a method and system for detecting engine blade deformation. The method of this invention determines whether the blade is deformed by comparing the ideal blade profile with the profile of the blade to be inspected, effectively ensuring accuracy, achieving a high degree of automation, and improving inspection efficiency. The system of this invention has a simple structure, enabling both high-efficiency inspection and high-precision detection of blade deformation.

[0007] 2. Technical Solution

[0008] To solve the above problems, the present invention adopts the following technical solution.

[0009] A method for detecting engine blade deformation includes the following steps:

[0010] S1: Acquire the original image, which includes the ball cap portion of the engine and the blade portion to be detected;

[0011] S2: Extract the outline of the spherical cap and the outline of the leaf to be detected from the original image respectively;

[0012] S3: Calculate the ideal blade profile of the blade to be detected based on the extracted spherical cap profile;

[0013] S4: Compare the ideal blade profile with the extracted blade profile to determine whether it has been deformed.

[0014] Furthermore, step S3 specifically includes the following steps:

[0015] S31: Calculate the intersection coordinates of the extracted cap contour. The intersection coordinates refer to the coordinates of the intersection of the major axis of the cap contour with the cap contour on the original image, and the coordinates of the intersection of the minor axis of the cap contour with the cap contour on the original image.

[0016] S32: Determine the transformation relationship between world coordinates and camera coordinates using the intersection coordinates from step S31;

[0017] S33: Extract the leaf contour adjacent to the leaf to be detected, which is called the reference leaf contour;

[0018] S34: Extract the inflection points in the reference blade profile, and solve the coordinate relationship between the inflection points and the camera based on the coordinate relationship between the inflection points and the spherical cap.

[0019] S35: Using the inflection point as a reference, the coordinate relationship between other blades and the camera is solved according to the engine blade model. Then, according to the pinhole camera imaging model, the undeformed blade contour is mapped onto the original image to obtain the ideal blade contour.

[0020] Furthermore, when extracting the contour of the cap from the original image, the lower edge of the cap is selected for contour extraction.

[0021] Furthermore, step S2, the extraction of the cap outline, specifically includes the following steps:

[0022] S21: Establish an HSV color model, convert the RGB of the original image to HSV, and extract the spherical cap region in the original image using the hue channel threshold.

[0023] S22: The image of the cap region is processed by erosion and dilation, and the edges of the region are extracted to obtain the outline of the circle at the root of the cap.

[0024] Furthermore, step S22 is followed by step S23: fitting the contour of the obtained ball cap root circle to obtain the fitted contour.

[0025] Furthermore, in step S1, the original image is acquired by mounting a camera on a robotic arm, which is mounted on an AGV trolley. The AGV trolley is controlled to automatically move along the engine intake duct to the corresponding position of the engine blades. After the robotic arm moves to the designated posture, the camera takes a picture to obtain the original image.

[0026] Furthermore, step S5 is included after step S4: performing linear fitting on the ideal blade profile and the extracted blade profile to be detected, and comparing it with the corresponding linear equation parameters to achieve quantitative judgment of deformation.

[0027] A system for detecting engine blade deformation using any of the methods described above, comprising:

[0028] Imaging unit: Used to capture raw images, which include the ball cap part of the engine and the blade part to be detected;

[0029] Extraction unit: used to extract the required regions from the original image according to the requirements;

[0030] Calculation unit: used to calculate the ideal blade profile of the blade to be inspected;

[0031] Comparison unit: Used to compare the ideal blade profile with the blade profile to be detected.

[0032] 3. Beneficial effects

[0033] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0034] (1) This invention acquires an original image of a blade with a cap and a blade to be tested, extracts the cap contour and the blade contour from the original image, calculates the ideal blade contour based on the cap contour, and then compares it with the blade contour to determine whether the blade is deformed. A reference standard is established, making the final judgment on the deformation of the blade to be tested more intuitive and accurate, thus improving the accuracy of the detection. The whole process is simple to operate and highly automated, effectively improving the problems of slow detection efficiency and low accuracy caused by manual visual inspection in the past.

[0035] (2) This invention processes the spherical cap contour. Since the projection of the spherical cap contour is elliptical, the major axis and minor axis of the ellipse are solved. The transformation relationship between world coordinates and camera coordinates is solved by using the four intersection points of the major and minor axes with the ellipse, thus ensuring accuracy. At the same time, the inflection point of the reference blade contour is extracted. Since the positional relationship of the inflection point relative to the spherical cap is determined, the coordinates of the inflection point are used as the reference to solve the coordinate relationship of other blades relative to the camera. The complete and ideal blade contour is mapped onto the original image. The entire process has high solution accuracy and can reduce the generation of errors to a certain extent.

[0036] (3) The present invention extracts the contour of the lower edge of the ball cap to further ensure the smooth progress of subsequent steps and avoid interference with subsequent steps; secondly, the original image is acquired by using a camera in conjunction with an AGV intelligent vehicle. The camera captures clear images and has a wide shooting range, and the intelligent vehicle can drive the camera to the corresponding position, thus ensuring that the acquisition of the original image is more convenient and faster.

[0037] (4) The system structure of the present invention is simple, which can realize both high-efficiency inspection of blade deformation and high-precision detection of blade deformation. Furthermore, the units are closely connected while not interfering with each other and working stably. At the same time, the system does not rely on manual inspection or sensor feedback. By processing the original image with images and algorithms, it can accurately and quickly complete the judgment of blade deformation, which greatly improves work efficiency and reduces labor and equipment costs. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the process of the present invention;

[0039] Figure 2 This is a schematic diagram of the original image obtained from the leaf detection. Detailed Implementation

[0040] The present invention will now be further described with reference to specific embodiments and accompanying drawings.

[0041] Example 1

[0042] like Figure 1 and Figure 2 As shown, a method for detecting engine blade deformation includes the following steps:

[0043] S1: Acquire the original image. The original image includes the spherical cap portion of the engine and the blade portion to be detected. It's worth noting that the original image contains not only the blade portion to be detected, but also adjacent blades. Specifically, in this step, to acquire the original image, a camera is mounted on a robotic arm, which is mounted on an AGV (Automated Guided Vehicle). The robotic arm's configuration facilitates camera adjustments to capture the required image. By controlling the AGV to automatically move along the engine intake duct to the corresponding position of the engine blade, and after the robotic arm reaches the designated posture, the camera captures the original image. Because the camera provides clear images with a wide shooting range and is easy to control, this step... Since the cap is a semi-ellipsoidal metal, the outline of its upper edge is not fixed when photographed from different angles. Therefore, a fixed outline with a relatively clear lower edge of the cap is chosen as a reference. The center of the circle at the lower edge of the cap is taken as the origin of the coordinate system, and the diameter of this circle is known. The plane containing the circle is the yz plane, and a coordinate system is established outwards from the x-axis perpendicular to the yz plane, serving as the world coordinate system. After the camera is mounted on the robotic arm of the AGV, the transformations from image coordinates to pixel coordinates and from camera coordinates to image coordinates are determined. Since the coordinate transformation relationship can be determined by camera calibration, and the transformation is not the focus of this application, and the transformation relationship is a conventional technique in this field, it will not be described in detail. Furthermore, the distortion parameters of the camera lens (which can be obtained from the original camera information) are also determined. Only the transformation from world coordinates to camera coordinates is unknown. If the position and orientation of the camera in the world coordinate system are known, the imaging position of any known point in the original image can be calculated according to the pinhole camera imaging model relationship.

[0044] S2: Extract the outline of the spherical cap and the outline of the leaf to be detected from the original image respectively; it should be noted in this step that since the projection of the angled circle in the image is an ellipse, the outline of the spherical cap in the original image is elliptical for a semi-ellipsoidal spherical cap; and when extracting the outline of the spherical cap from the original image, the lower edge of the spherical cap is selected for outline extraction; specifically, the extraction of the spherical cap outline includes the following steps:

[0045] S21: Establish an HSV color model, convert the RGB of the original image to HSV, use color to distinguish contours, ensure the accuracy of extraction, and reduce errors; use the hue channel threshold to extract the spherical cap region from the original image.

[0046] S22: The image of the spherical cap region is processed by erosion and dilation, and the edges of the region are extracted to obtain the outline of the circle at the root of the spherical cap, that is, the elliptical outline of the lower edge of the spherical cap.

[0047] S23: Fit the outline of the obtained circular base of the sphere to obtain the fitted elliptical outline. Step S23 is set to obtain more accurate elliptical information, providing a precise guarantee for subsequent calculations and processing.

[0048] S3: Based on the extracted spherical cap contour, convert and calculate the ideal blade contour of the blade to be detected; step S3 specifically includes the following steps:

[0049] S31: Calculate the intersection coordinates of the extracted elliptical contour, i.e., the fitted elliptical contour. The intersection coordinates refer to the coordinates of the intersection of the major axis of the elliptical contour with the elliptical contour on the original image, and the coordinates of the intersection of the minor axis of the elliptical contour with the elliptical contour on the original image.

[0050] S32: Determine the transformation relationship between world coordinates and camera coordinates using the intersection coordinates from step S31;

[0051] S33: Extract the blade profile adjacent to the blade to be detected, which is called the reference blade profile. Since the root of the spherical cap is circular and is a symmetrical figure about the center, the circular figure can only determine the coordinates of the center of the extracted circle in the camera coordinate system and the direction of the axis of the spherical cap. The purpose of extracting the reference blade profile in this step is to determine the angle at which the engine blade stops, so as to have a better and more specific understanding of the engine blade.

[0052] S34: Extract the inflection points in the reference blade profile, and solve the coordinate relationship between the inflection points and the camera based on the coordinate relationship between the inflection points and the spherical cap; the inflection point of the blade profile refers to the corner of the profile. For example, for a rectangular blade, the inflection point is the corner of the rectangle.

[0053] S35: Using the inflection point as a reference, the coordinate relationship between other blades and the camera is solved according to the engine blade model. Then, according to the pinhole camera imaging model, the undeformed blade contour is mapped onto the original image to obtain the ideal blade contour.

[0054] S4: Compare the ideal blade profile with the extracted blade profile to determine whether it has been deformed.

[0055] Step S4 is followed by step S5: Linear fitting is performed on the ideal blade profile and the extracted blade profile to be detected, and the results are compared with the corresponding linear equation parameters to achieve a quantitative judgment of deformation. Specifically: the linear equation is A*x+B*y+C=0; the linear equation of the ideal model is A0*x+B0*y+C0=0; E=(A-A0)*(A-A0)+(B-B0)*(B-B0)+(C-C0)*(C-C0) is used as the standard for judging whether the blade is deformed. When E>E b The blade under test is considered to be deformed, E b These are experience points.

[0056] This invention acquires an original image containing a spherical cap and a blade to be inspected. The spherical cap contour and the blade contour are extracted from the original image. An ideal blade contour is then calculated based on the spherical cap contour and compared with the actual blade contour to determine if the blade is deformed. This establishes a reference standard, making the final determination of blade deformation more intuitive and accurate, thus improving detection precision. The entire process is simple to operate and highly automated, effectively overcoming the problems of slow efficiency and low accuracy associated with traditional manual visual inspection.

[0057] Example 2

[0058] A system for detecting engine blade deformation using a method as described in the above embodiments includes:

[0059] Imaging unit: Used to capture raw images, which include the ball cap part of the engine and the blade part to be detected;

[0060] Extraction unit: used to extract the required regions from the original image according to the requirements;

[0061] Calculation unit: used to calculate the ideal blade profile of the blade to be inspected;

[0062] Comparison unit: Used to compare the ideal blade profile with the blade profile to be detected.

[0063] The system structure of this invention is simple, enabling both high-efficiency inspection and high-precision detection of blade deformation. Furthermore, while the units are tightly connected, they do not interfere with each other, ensuring stable operation. Moreover, this system does not rely on manual inspection or sensor feedback; by processing the original images using image and algorithmic methods, it can accurately and quickly determine blade deformation, greatly improving work efficiency while reducing labor and equipment costs.

[0064] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

Claims

1. A method for detecting engine blade deformation, characterized in that: Includes the following steps: S1: Acquire the original image, which includes the ball cap portion of the engine and the blade portion to be detected; S2: Extract the outline of the spherical cap and the outline of the leaf to be detected from the original image respectively; S3: Calculate the ideal blade profile of the blade to be detected based on the extracted spherical cap profile; S4: Compare the ideal blade profile with the extracted blade profile to determine whether it has been deformed; Step S3 specifically includes the following steps: S31: Calculate the intersection coordinates of the extracted cap contour. The intersection coordinates refer to the coordinates of the intersection of the major axis of the cap contour with the cap contour on the original image, and the coordinates of the intersection of the minor axis of the cap contour with the cap contour on the original image. S32: Determine the transformation relationship between world coordinates and camera coordinates using the intersection coordinates from step S31; S33: Extract the leaf contour adjacent to the leaf to be detected, which is called the reference leaf contour; S34: Extract the inflection points in the reference blade profile, and solve the coordinate relationship between the inflection points and the camera based on the coordinate relationship between the inflection points and the spherical cap. S35: Using the inflection point as a reference, the coordinate relationship between other blades and the camera is solved according to the engine blade model. Then, according to the pinhole camera imaging model, the undeformed blade contour is mapped onto the original image to obtain the ideal blade contour.

2. The method for detecting engine blade deformation according to claim 1, characterized in that: When extracting the contour of the cap from the original image, select the lower edge of the cap for contour extraction.

3. The method for detecting engine blade deformation according to claim 2, characterized in that: Step S2, which involves extracting the outline of the spherical cap, specifically includes the following steps: S21: Establish an HSV color model, convert the RGB of the original image to HSV, and extract the spherical cap region in the original image using the hue channel threshold. S22: The image of the cap region is processed by erosion and dilation, and the edges of the region are extracted to obtain the outline of the circle at the root of the cap.

4. The method for detecting engine blade deformation according to claim 3, characterized in that: Step S22 is followed by step S23: fitting the outline of the obtained ball cap root circle to obtain the fitted outline.

5. The method for detecting engine blade deformation according to claim 1, characterized in that: In step S1, the original image is acquired by mounting a camera on a robotic arm, which is then mounted on an AGV (Automated Guided Vehicle). The AGV is controlled to automatically move along the engine intake duct to the corresponding position on the engine blade. After the robotic arm reaches the designated posture, the camera captures the original image.

6. The method for detecting engine blade deformation according to claim 1, characterized in that: Step S4 is followed by step S5: performing linear fitting on the ideal blade profile and the extracted blade profile to be detected, and comparing it with the corresponding linear equation parameters to achieve quantitative judgment of deformation.

7. A system for applying the method for detecting engine blade deformation as described in any one of claims 1-6, characterized in that: include: Imaging unit: Used to capture raw images, which include the ball cap part of the engine and the blade part to be detected; Extraction unit: used to extract the required regions from the original image according to the requirements; Calculation unit: used to calculate the ideal blade profile of the blade to be inspected; Comparison unit: Used to compare the ideal blade profile with the blade profile to be detected.

Images