Full-automatic inspection method, system and equipment for high-pressure turbine blade of aero-engine

Abstract

The invention belongs to the field of aero-engine blade detection, particularly relates to a full-automatic inspection method, system and equipment for aero-engine high-pressure turbine blades, and aims to solve the problems that in the prior art, a traditional borescope inspection method is difficult to align, the 360-degree damage condition of each blade is difficult to comprehensively judge, and the detection precision is low. The method comprises the following steps: acquiring image data of a to-be-detected blade, inputting the image data into a trained blade damage identification model to obtain a prediction result, acquiring key feature points of a blade edge in the image data, and calculating a pixel distance of the key feature points on the image data according to a pre-constructed blade radian model, calculating the conversion proportion between the size and the actual size; and calculating the size of the bounding box based on the bounding box coordinates of the damage position, and obtaining the actual size of the damage of the to-be-detected blade in combination with the conversion proportion. According to the method, the images shot at different angles can be quickly and accurately aligned to the same blade, so that the time required for inspection is shortened, and the detection precision is improved.

Description

Background Art

[0002] The efficient operation and safety of aviation turbofan engines are inseparable from the strict monitoring of its key components, the High Pressure Turbine (HPT) blades. HPT blades work in extreme environments, and are subjected to multiple tests of high temperature, high pressure, and high-speed rotation, making them prone to damage such as cracks, corrosion, and wear. Therefore, regular and accurate damage detection and evaluation of HPT blades is essential to ensure the reliability of the engine and extend its service life.

[0003] The traditional inspection method of high-pressure turbine blades mainly relies on borescope inspection technology, that is, visual inspection of the blades from different holes of the engine (such as leading edge and trailing edge holes) through equipment such as endoscopes. This method has the following main problems:

[0004] Difficult alignment: Since the blade itself may not have obvious feature marks, it is difficult to quickly and accurately align the images taken through the leading edge and trailing edge holes to the same blade. This makes it impossible to comprehensively judge the damage of each blade within a 360-degree range, increasing the complexity and uncertainty of the inspection.

[0005] Incomplete records: Traditional borescope inspections usually only record the most severely damaged blades, and fail to fully track the damage of each blade. Although this selective recording method can save time and resources, it may also cause some potential minor damage to be overlooked, thus affecting the accurate assessment of the overall health of the entire high-pressure turbine blade group.

[0006] Low accuracy: Traditional methods are highly dependent on the experience and skills of inspectors. There may be large differences in judgment between different personnel, which not only affects the consistency of the test results, but also makes the testing process lack standardization, making it difficult to ensure that each test can meet the same quality standards.

[0007] Based on this, the present invention proposes a fully automatic inspection method, system and equipment for high-pressure turbine blades of an aircraft engine. Summary of the invention

[0008] In order to solve the above-mentioned problems in the prior art, namely, in the prior art, the traditional borescope inspection method has the problems of alignment difficulties and incomplete records, making it difficult to comprehensively judge the 360-degree damage condition of each blade, and being unable to comprehensively track the damage data of all blades, as well as the problem of low detection accuracy, the present invention provides a fully automatic inspection method, system and equipment for aircraft engine high-pressure turbine blades.

[0009] In a first aspect of the present invention, a fully automatic inspection method for a high-pressure turbine blade of an aircraft engine is provided, the method comprising:

[0010] Acquire image data of the blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes the damage type and the bounding box coordinates of the damage location; wherein the image data is obtained by acquiring images of different positions of the blade based on different probes and performing image alignment;

[0011] Acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size according to the pixel distance;

[0012] The size of the bounding box is calculated based on the bounding box coordinates of the damage location, and then combined with the conversion ratio to obtain the actual size of the blade damage to be detected;

[0013] Compare the damage location and actual size with the standards in the aircraft maintenance manual, and output an inspection report and repair plan.

[0014] In some preferred embodiments, the damage identification model includes a basic feature extraction network, a multi-scale feature map generation network, and a prediction network that are connected in sequence.

[0015] In some preferred embodiments, the damage identification model, the training method thereof comprises the following steps:

[0016] Step S1, inputting the data set into the damage identification model to obtain the prediction result;

[0017] Step S2, calculating the total loss of the prediction result and the actual annotation based on the loss function;

[0018] Step S3, calculating the gradient through the back propagation algorithm, and using the optimizer to update the weight of the damage identification model so that the loss function value gradually decreases;

[0019] Step S4, jump to step S1, and execute steps S1 to S3 until the loss function converges or reaches a predetermined number of training rounds, and then stop jumping.

[0020] In some preferred embodiments, the loss function includes classification loss, positioning loss and confidence loss.

[0021] In some preferred embodiments, the pixel distance of the key feature points on the image data is calculated according to the pre-built blade radian model, and the method is:

[0022] Obtain a standard object of known size and take images of the standard object from multiple angles and distances using the same camera settings and environmental conditions as the blades.

[0023] Extract the arc position of the standard object and calculate the image pixel distances at different arc positions;

[0024] A conversion relationship between pixel size and actual size is established based on the actual size of the standard object and the pixel distance in the image;

[0025] The pixel distance of the key feature points on the image data is calculated based on the transformation relationship.

[0026] In some preferred implementations, the conversion ratio between the pixel distance and the actual size is calculated according to the method as follows:

[0027] The actual size of the key feature point on the actual leaf is obtained, and the ratio of the actual size to the pixel distance is used as the conversion ratio.

[0028] In some preferred embodiments, the image data is acquired based on two probes arranged at the leading edge and trailing edge of the HPT1 level blades. The two probes are used to collect images of the leading edge, trailing edge, blade tip and blade root of the blades from multiple perspectives, and after image alignment, 360-degree image information of each blade is obtained as image data.

[0029] In some preferred embodiments, the image alignment is performed by:

[0030] The rotation speed and rotation amount signals transmitted by the engine rotation tool are obtained, the actual movement amount of the HPT rotor blade per unit time is calculated, and the time for the blade to pass through the two probes is obtained respectively, so as to obtain the pictures of the blade taken by different probes, and multiple pictures are associated with the same blade to obtain image data at multiple angles on the same blade.

[0031] Another aspect of the present invention provides a fully automatic inspection system for high-pressure turbine blades of an aircraft engine. Based on a fully automatic inspection method for high-pressure turbine blades of an aircraft engine, the system comprises:

[0032] A damage acquisition module is configured to acquire image data of a blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes a damage type and a bounding box coordinate of a damage location;

[0033] A conversion ratio acquisition module, configured to acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size;

[0034] A damage actual size acquisition module, which is configured to calculate the size of the bounding box based on the bounding box coordinates of the damage location, and then obtain the actual size of the blade damage to be detected in combination with the conversion ratio;

[0035] The maintenance plan generating module is configured to compare the damage location and the actual size of the damage with the standards in the aircraft maintenance manual, and output an inspection report and a maintenance plan.

[0036] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0037] at least one processor; and

[0038] a memory communicatively connected to at least one of the processors; wherein,

[0039] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement a fully automatic inspection method for high-pressure turbine blades of an aircraft engine.

[0040] Beneficial effects of the present invention:

[0041] Improve inspection accuracy: By using a trained blade damage recognition model, the location, type and bounding box coordinates of blade damage can be accurately identified, greatly improving the accuracy and reliability of damage detection. In addition, the actual size of the damage is calculated based on the bounding box coordinates of the damage location, further ensuring the accuracy of damage assessment.

[0042] Enhanced inspection efficiency: The fully automated inspection process reduces the need for manual operation and improves the speed and efficiency of inspection. In particular, the application of image alignment technology allows images taken at different angles to be quickly and accurately aligned to the same blade even without obvious feature identification, enabling 360-degree all-round inspection, greatly shortening the time required for inspection.

[0043] Comprehensive tracking: The method of the present invention not only focuses on the blades with the most serious damage, but also comprehensively tracks the damage of each blade, which helps to more accurately assess the overall health of the entire high-pressure turbine blade group. This is conducive to early detection of potential minor damage, avoiding small problems from turning into major failures, and effectively extending the service life of the engine.

[0044] Optimize maintenance decisions: The maintenance plan generated based on the actual size of the blade damage to be detected provides maintenance personnel with more scientific and reasonable guidance, helps to arrange maintenance plans reasonably, reduces maintenance costs, and ensures the safe operation of the engine.

[0045] Reduce human errors: The automated inspection process reduces the impact of human factors, avoids errors caused by personal experience and skill differences, and ensures the consistency and reliability of inspection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0047] Figure 1 It is a schematic flow chart of a fully automatic inspection method for high-pressure turbine blades of an aircraft engine according to the present invention;

[0048] Figure 2 This is a front view of a fully automatic inspection device for high-pressure turbine blades of an aircraft engine according to the present invention;

[0049] Figure 3 It is a side view of a fully automatic inspection device for high-pressure turbine blades of an aircraft engine according to the present invention;

[0050] Figure 4 It is a structural schematic diagram of a probe in a fully automatic inspection device for high-pressure turbine blades of an aircraft engine according to the present invention;

[0051] Figure 5 It is a schematic diagram of a terminal in a fully automatic inspection device for high-pressure turbine blades of an aircraft engine of the present invention;

[0052] Figure 6 It is a multi-angle schematic diagram of two probes in a fully automatic inspection device for high-pressure turbine blades of an aircraft engine of the present invention;

[0053] Figure 7 It is a structural diagram of a computer system of a server for implementing the method, system, and device embodiments of the present application. DETAILED DESCRIPTION

[0054] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0055] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0056] The present invention provides a fully automatic inspection method for aeroengine high-pressure turbine blades, the method comprising:

[0057] Acquire image data of the blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes the damage type and the bounding box coordinates of the damage location; wherein the image data is obtained by acquiring images of different positions of the blade based on different probes and performing image alignment;

[0058] Acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size according to the pixel distance;

[0059] The size of the bounding box is calculated based on the bounding box coordinates of the damage location, and then combined with the conversion ratio to obtain the actual size of the blade damage to be detected;

[0060] Compare the damage location and actual size with the standards in the aircraft maintenance manual, and output an inspection report and repair plan.

[0061] In order to more clearly explain the fully automatic inspection method of the high-pressure turbine blade of an aircraft engine of the present invention, the following is combined with Figure 1 Each step in the embodiment of the present invention is described in detail as follows:

[0062] Acquire image data of the blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes the damage type and the bounding box coordinates of the damage location; wherein the image data is obtained by acquiring images of different positions of the blade based on different probes and performing image alignment;

[0063] The image data described in the present invention is obtained based on two probes set at the leading edge and trailing edge of the HPT1 level blade. The two probes are used to collect images of the leading edge, trailing edge, blade tip and blade root of the blade from multiple perspectives, and after image alignment, 360-degree image information of each blade is obtained as image data.

[0064] The image alignment method is as follows:

[0065] The rotation speed and rotation amount signals transmitted by the engine rotation tool are obtained, the actual movement amount of the HPT rotor blade per unit time is calculated, and the time for the blade to pass through the two probes is obtained respectively, so as to obtain the pictures of the blade taken by different probes, and multiple pictures are associated with the same blade to obtain image data at multiple angles on the same blade.

[0066] Specifically, after calculating the actual movement of the HPT rotor blades per unit time, we can get the time when blade 1 passes probe A and probe B, and obtain the pictures captured by blade 1 in different lenses. These pictures are then associated with blade 1 to obtain the image information of blade 1 at different angles. The image data acquisition methods for other blades are as described above and will not be repeated here.

[0067] In the present invention, the damage identification model includes a basic feature extraction network, a multi-scale feature map generation network and a prediction network which are connected in sequence.

[0068] In this embodiment, the damage identification model is built based on YOLOv5, and its architecture includes three parts: Backbone, Neck and Head:

[0069] Backbone: used to extract basic image features, usually using a convolutional neural network (CNN) structure, such as CSPDarknet.

[0070] Neck: used to generate multi-scale feature maps, usually using FPN (Feature Pyramid Network) or PANet (Path Aggregation Network) structure.

[0071] Head: Used to predict the category, location, and confidence of each target object.

[0072] The damage identification model in the present invention has a training method comprising the following steps:

[0073] Step S1, inputting the data set into the damage identification model to obtain the prediction result;

[0074] Step S2, calculating the total loss of the prediction result and the actual annotation based on the loss function;

[0075] Step S3, calculating the gradient through the back propagation algorithm, and using the optimizer to update the weight of the damage identification model so that the loss function value gradually decreases;

[0076] Step S4, jump to step S1, and execute steps S1 to S3 until the loss function converges or reaches a predetermined number of training rounds, and then stop jumping.

[0077] The dataset is collected through daily engine borescope inspections, and contains a large number of image datasets with target annotations of HPT blade damage (including coating peeling, ablation, cracks, burn-through, material loss, etc.), and the annotations include the category and bounding box of the target object.

[0078] Format the images and annotation data in the dataset, such as resizing the images, normalizing the pixel values, etc. At the same time, data augmentation operations are performed, such as random cropping, rotation, scaling, color adjustment, etc., to increase data diversity and prevent overfitting.

[0079] The optimizer may be Adam or SGD.

[0080] The loss function in this embodiment includes classification loss, positioning loss and confidence loss.

[0081] Classification loss: measures the difference between the predicted target category and the actual category, usually using cross entropy loss.

[0082] Positioning loss: measures the difference between the predicted bounding box and the actual bounding box, usually using IoU (Intersection over Union) or GIoU (Generalized IoU) loss.

[0083] Confidence loss: measures the difference between the predicted probability of the target existence and the actual situation.

[0084] After the training of the damage identification model of the present invention, the model evaluation and testing are carried out, and the method is as follows:

[0085] Validation set evaluation: Use an independent validation data set to evaluate model performance. Common indicators include mAP (meanAverage Precision).

[0086] Hyperparameter adjustment: According to the evaluation results of the validation set, adjust the model hyperparameters (such as learning rate, batch size, etc.) to further optimize the model performance.

[0087] Test set evaluation: Use the test dataset for final evaluation to ensure that the model performs well on unseen data.

[0088] Model export: Export the trained model to a deployable format such as ONNX, TorchScript, etc.

[0089] Inference: Use the model in real-world applications for real-time or batch object detection.

[0090] The HPT blade damage recognition model is trained through the above steps.

[0091] Acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size according to the pixel distance;

[0092] In the present invention, the trained YOLOv5 model is used to detect the damage in the HPT leaf image, the bounding box coordinates of each damage are obtained, and the coordinates of the key feature points of the leaf edge are detected by manual annotation.

[0093] In order to accurately calculate the actual size of the damage, it is necessary to establish a blade camber model. This can be achieved through the following methods:

[0094] Calibration image: Use a calibration image of known size to obtain the image pixel distances of the blade at different arc positions.

[0095] Geometric modeling: Establish a mathematical model of the blade surface based on the design parameters and geometric shape of the blade.

[0096] Based on the method of calibrating the image, the pixel distance is calculated. Specifically, the pixel distance of the key feature points on the image data is calculated according to the pre-built blade curvature model. The method is:

[0097] Obtain a standard object of known size and take images of the standard object from multiple angles and distances using the same camera settings and environmental conditions as the blades.

[0098] Extract the arc position of the standard object and calculate the image pixel distances at different arc positions;

[0099] A conversion relationship between pixel size and actual size is established based on the actual size of the standard object and the pixel distance in the image;

[0100] The pixel distance of the key feature points on the image data is calculated based on the transformation relationship.

[0101] Based on the geometric modeling method, the specific method is:

[0102] Collect blade design parameters: Obtain blade design drawings and technical parameters from the manufacturer, including but not limited to blade length, width, thickness, curvature radius, etc.

[0103] Build a 3D model: Use CAD software (such as SolidWorks, AutoCAD, etc.) to build a 3D geometric model of the blade according to the design parameters. The model should reflect the actual shape and size of the blade as accurately as possible, especially the curved part of the blade.

[0104] Model verification: The accuracy of the model is verified by actually measuring the key dimensions of the blade and comparing them with the 3D model. This step is very important to ensure the accuracy of subsequent dimensional calculations.

[0105] Model application: Import the established 3D model into the image processing software, and use the geometric information provided by the model to assist in calculating the actual distance between any two points on the blade, especially the distance calculation on the curved surface.

[0106] In the present invention, the conversion ratio between the pixel distance and the actual size is calculated according to the method as follows:

[0107] The actual size of the key feature point on the actual leaf is obtained, and the ratio of the actual size to the pixel distance is used as the conversion ratio.

[0108] In the present invention, by calculating the image pixel distance of the arc position of the standard object, the projection of the arc position in the plane can be obtained, and by calculating the conversion relationship, the real arc length or area can be obtained.

[0109] The size of the bounding box is calculated based on the bounding box coordinates of the damage location, and then combined with the conversion ratio to obtain the actual size of the blade damage to be detected;

[0110] In the present invention, the actual size of the damage is calculated using the above ratio according to the pixel size of the detected damage boundary box, and the influence of the arc of the damage location on the size calculation is considered, so as to obtain an accurate damage size.

[0111] Compare the damage location and actual size with the standards in the aircraft maintenance manual, and output an inspection report and repair plan.

[0112] Specifically, based on the damage location and actual size of the damage in the test results, the corresponding standards (inspection area, damage location, damage size) in the aircraft's AMM (Aircraft Maintenance Manual) are compared to output corresponding measures, such as monitoring the damage or directly removing the engine for repair.

[0113] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0114] The second embodiment of the present invention provides a fully automatic inspection system for high-pressure turbine blades of an aircraft engine. Based on a fully automatic inspection method for high-pressure turbine blades of an aircraft engine, the system includes:

[0115] A damage acquisition module is configured to acquire image data of a blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes a damage type and a bounding box coordinate of a damage location;

[0116] A conversion ratio acquisition module, configured to acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size;

[0117] A damage actual size acquisition module, which is configured to calculate the size of the bounding box based on the bounding box coordinates of the damage location, and then obtain the actual size of the blade damage to be detected in combination with the conversion ratio;

[0118] The maintenance plan generating module is configured to compare the damage location and the actual size of the damage with the standards in the aircraft maintenance manual, and output an inspection report and a maintenance plan.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0120] It should be noted that the above embodiment provides a fully automatic inspection system for high-pressure turbine blades of an aircraft engine, and only uses the division of the above functional modules as an example. In practical applications, the above functional distribution can be completed by different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiments can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not regarded as improper limitations on the present invention.

[0121] An electronic device according to a third embodiment of the present invention includes:

[0122] at least one processor; and

[0123] a memory communicatively connected to at least one of the processors; wherein,

[0124] The memory stores instructions that can be executed by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned fully automatic inspection method for high-pressure turbine blades of an aircraft engine.

[0125] A computer-readable storage medium according to a fourth embodiment of the present invention stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned fully automatic inspection method for high-pressure turbine blades of an aircraft engine.

[0126] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the storage device and processing device described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0127] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0128] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system of a server for implementing the method, system, and device embodiments of the present application. Figure 7 The server shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0129] like Figure 7 As shown, the computer system includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 702 or the program loaded from the storage part 708 to the random access memory (RAM) 703. In the RAM 703, various programs and data required for system operation are also stored. The CPU 701, the ROM 702 and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0130] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed so that a computer program read therefrom is installed into the storage section 708 as needed.

[0131] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the method of the present application are executed. It should be noted that the above-mentioned computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, - but not limited to - a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, an apparatus or a device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, an apparatus or a device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0132] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0133] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0134] A fifth embodiment of the present invention, as Figure 2-Figure 6 As shown, a fully automatic inspection device for high-pressure turbine blades of an aircraft engine is proposed, comprising a rotation drive device, a first probe 1, a second probe 2 and a terminal 3;

[0135] The rotation drive device is drivingly connected to the rotation shaft of the engine high-pressure rotor, and is used to drive the rotation shaft to rotate;

[0136] The first probe 1 passes through the engine outer casing 4 and the engine inner casing 5 and is arranged on one side of the leading edge of the blade 6, for acquiring image data of the leading edge of the blade 6;

[0137] The second probe 2 passes through the engine outer casing 4 and the engine inner casing 5 and is arranged on one side of the trailing edge of the blade 6, so as to obtain image data of the trailing edge of the blade 6;

[0138] The rotation drive device, the first probe 1 and the second probe 2 are all electrically connected to the terminal 3, and the terminal 3 is used to control the rotation speed, rotation direction and rotation amount of the rotation drive device, and obtain image data of the leading edge and the trailing edge, and obtain the damage area in combination with the inspection method.

[0139] Among them, Figure 5 As shown, the terminal 3 includes a calculation part and an interactive PAD31. The terminal 3 can control the rotation speed, rotation direction and rotation amount of the rotation drive device; and receive the information of the rotation speed, rotation direction and rotation amount of the rotation tool, so as to accurately locate each blade 6; the terminal receives the images of the first probe 1 and the second probe 2, and the rotation tool has the functions of image recognition and size calculation by calculating. When the blade 6 passes through the middle position of the probe, the image of the blade 6 is captured, and the blade 6 in the image is located by data such as the rotation amount; the damage on the engine blade is identified by the damage recognition model, and the actual size and specific position of the damage are calculated by analyzing the pixel size of the damage in the image and comparing the ratio of the actual size of the blade edge to the pixel size. At the same time, for the front edge area of ​​the damaged blade with overlapping areas in the observation image, the comparison relationship of the two cameras to the unified position can also be used to form the measurement conditions of the binoculars, thereby providing more accurate measurement. In addition, this part can also write engine reports, call the standards in the AMM manual, and perform automatic comparison. Determine the availability of damage; and provide damage comparison between the two inspections before and after.

[0140] in addition, Figure 5 Also included is a second-stage high-pressure turbine cover ring 13 and a second-stage high-pressure turbine rotor blade 14, which are prior arts and will not be described in detail here.

[0141] For further explanation of the present invention, see Figure 2 and Figure 3 The first probe 1 includes a first probe body 11 and a lens portion 12 at the end of the first probe;

[0142] The first probe body 11 and the lens portion 12 at the end of the first probe are connected via a rotating mechanism, and the rotating mechanism is used to drive the lens portion 12 at the end of the first probe to rotate to an angle parallel to the length direction of the HPT1 level guide 7;

[0143] The first probe tube body 11 and the lens portion 12 at the end of the first probe enter the interior of the engine outer casing 4 from the probe port of the HPT1 stage guide 7 .

[0144] Specifically, in the initial state, the rotating mechanism does not drive the lens part 12 at the end of the first probe to rotate. At this time, there is no angle between the first probe tube body 11 and the lens part 12 at the end of the first probe. At this time, the lens part 12 at the end of the first probe enters the engine outer casing 4 from the probe port of the HPT1 stage guide 7, and after passing through the engine inner casing 5, the rotating device is started to drive the lens part 12 at the end of the first probe to deflect backward by 40 degrees to approach the leading edge of the blade 6 to obtain a better observation field of view. During the inspection, the leading edge of the HPT1 stage rotor blade 6 is observed from the front.

[0145] For further explanation of the present invention, see Figure 2 and Figure 3 The second probe 2 includes a second probe body 21 and a lens portion 22 at the end of the second probe;

[0146] The second probe tube body 21 and the lens part 22 at the end of the second probe enter the interior of the engine outer casing 4 from the probe port of the HPT2 stage guide 8, the length direction of the second probe tube body 21 is perpendicular to the length direction of the engine outer casing 4, the second probe tube body 21 and the lens part 22 at the end of the second probe are fixed, and the length direction of the second probe tube body 21 and the length direction of the lens part 22 at the end of the second probe are arranged in parallel.

[0147] A rotating mechanism may be installed between the second probe body 21 and the lens portion 22 at the end of the second probe, or the second probe body 21 and the lens portion 22 may be directly fixed as a whole.

[0148] Among them, Figure 6 As shown, Figure 6 a is a schematic diagram of the leading edge at the first observation angle, b is a schematic diagram of the leading edge at the second observation angle, c is a schematic diagram of the trailing edge at the first observation angle, and d is a schematic diagram of the trailing edge at the second observation angle.

[0149] Figure 6 The first probe 1 and the second probe 2 in the figure are both schematic top views.

[0150] For further explanation of the present invention, see Figure 4 The lens part 12 at the end of the first probe and the lens part 22 at the end of the second probe both include multiple lenses and multiple light sources.

[0151] Wherein, the rotating mechanism is connected to the terminal via wireless communication.

[0152] Among them, the lenses and light sources in the lens part 12 at the end of the first probe and the lens part 22 at the end of the second probe are connected to the terminal through a wireless transmission module, which is embedded in the lens part 12 at the end of the first probe and the lens part 22 at the end of the second probe.

[0153] The wireless transmission module in the present invention can also be replaced with a line connection according to actual needs. The present invention does not specifically limit this content, and any method that can transmit data is within the protection scope of the present invention.

[0154] In this embodiment, each probe is equipped with two high-definition cameras with a diameter of less than 8 mm and a resolution of not less than 5 million pixels and two LED light sources. These cameras can respectively inspect the leading edge, blade basin area and blade back area of ​​the same blade 6, and provide a comprehensive high-definition inspection image of the blade 6. Compared with the inspection of the endoscope personnel, it can avoid the missed inspection of damage caused by poor observation angles.

[0155] The two lenses are respectively a first lens 121 and a second lens 122, and the light source includes a first light source 123 and a second light source 124;

[0156] The first light source 123 is disposed between the first lens 121 and the second lens 122 , and the second light source 124 is below the second lens 122 .

[0157] The rotation drive device includes a drive end and a connector;

[0158] The driving end is coaxially connected to the input shaft of the connector, the output shaft of the connector is coaxially connected to the rotating shaft, and the driving end is used to drive the rotating shaft to rotate.

[0159] The driving end in this embodiment can be controlled by a computing terminal or manually:

[0160] When the driving end is electric, a stepper motor can be selected, the stepper motor is connected to a power supply, the power supply is used to power the stepper motor, a wireless transmission module is installed on the stepper motor, and the stepper motor is connected to the terminal 3 through the wireless transmission module.

[0161] This part receives signals from the calculation and control part to control the speed and rotation amount of the stepper motor, thereby accurately controlling the rotation of the engine's high-pressure turbine shaft.

[0162] When the driving end is manual, a handle can be selected to switch to manual control. The handle also inputs electrical signals to control the motor to rotate, rather than using manpower to rotate the engine.

[0163] Wherein, the rotating shaft of the motor and the handle are detachably fixed.

[0164] Among them, the driving end has two signal input sources, one is the computing terminal, which can automatically control the driving end, so as to accurately control the rotation of the required inspection blade 6. However, this driving end can also be controlled by its own control handle, including controlling the rotation direction, rotation speed, rotation amount, etc. The control priority of the handle is higher than that of the computing terminal. The actual rotation amount, rotation speed and rotation direction information are collected by the sensor next to the motor shaft and fed back to the computing terminal.

[0165] See also Figure 2 The first probe tube body 11 and the second probe tube body 21 are both equipped with a limiting device 9 , and the limiting device 9 is arranged outside the engine outer casing 4 and overlaps with the probe opening opened on the engine outer casing 4 .

[0166] The limiting device 9 in the present invention can be selected as a limiting ring, and the diameter of the limiting ring is larger than the diameter of the probe port.

[0167] The above contents can keep the probe stuck on the engine to prevent the probe from falling.

[0168] As a further explanation of the present invention, a plurality of cameras are used to collect image data of the leading edge, the trailing edge, the blade tip and the blade root of the blade 6 .

[0169] The terms "first", "second", etc. are used to distinguish similar objects rather than to describe or indicate a particular order or sequence.

[0170] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.

[0171] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A fully automatic inspection method for high-pressure turbine blades of an aircraft engine, characterized in that: The method includes: Acquire image data of the blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes the damage type and the bounding box coordinates of the damage location; wherein the image data is obtained by acquiring images of different positions of the blade based on different probes and performing image alignment; Acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size according to the pixel distance; The size of the bounding box is calculated based on the bounding box coordinates of the damage location, and then combined with the conversion ratio to obtain the actual size of the blade damage to be detected; Compare the damage location and actual size with the standards in the aircraft maintenance manual, and output an inspection report and repair plan.

2. A fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 1, characterized in that: The damage identification model includes a basic feature extraction network, a multi-scale feature map generation network and a prediction network which are connected in sequence.

3. The fully automatic inspection method for high-pressure turbine blades of an aircraft engine according to claim 1, characterized in that: The damage identification model, and its training method comprises the following steps: Step S1, inputting the data set into the damage identification model to obtain the prediction result; Step S2, calculating the total loss of the prediction result and the actual annotation based on the loss function; Step S3, calculating the gradient through the back propagation algorithm, and using the optimizer to update the weight of the damage identification model so that the loss function value gradually decreases; Step S4, jump to step S1, and execute steps S1 to S3 until the loss function converges or reaches a predetermined number of training rounds, and then stop jumping.

4. A fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 3, characterized in that: The loss function includes classification loss, positioning loss and confidence loss.

5. The fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 1, characterized in that: The pixel distance of the key feature points on the image data is calculated based on the pre-built blade curvature model as follows: Obtain a standard object of known size and take images of the standard object from multiple angles and distances using the same camera settings and environmental conditions as the blades. Extract the arc position of the standard object and calculate the image pixel distances at different arc positions; A conversion relationship between pixel size and actual size is established based on the actual size of the standard object and the pixel distance in the image; The pixel distance of the key feature points on the image data is calculated based on the transformation relationship.

6. The fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 1, characterized in that: The conversion ratio between the pixel distance and the actual size is calculated as follows: The actual size of the key feature point on the actual leaf is obtained, and the ratio of the actual size to the pixel distance is used as the conversion ratio.

7. The fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 1, characterized in that: The image data is acquired based on two probes set at the leading edge and trailing edge of the HPT1 blade. The two probes are used to collect images of the leading edge, trailing edge, blade tip and blade root of the blade from multiple perspectives, and after image alignment, 360-degree image information of each blade is obtained as image data.

8. A fully automatic inspection method for aircraft engine high pressure turbine blades according to claim 7, characterized in that: The image alignment method is as follows: The rotation speed and rotation amount signals transmitted by the engine rotation tool are obtained, the actual movement amount of the HPT rotor blade per unit time is calculated, and the time for the blade to pass through the two probes is obtained respectively, so as to obtain the pictures of the blade taken by different probes, and multiple pictures are associated with the same blade to obtain image data at multiple angles on the same blade.

9. A fully automatic inspection system for high-pressure turbine blades of an aircraft engine, based on a fully automatic inspection method for high-pressure turbine blades of an aircraft engine according to any one of claims 1 to 8, characterized in that: The system includes: A damage acquisition module is configured to acquire image data of a blade to be detected, and input the image data as input data into a trained blade damage recognition model to obtain a prediction result, wherein the prediction result includes a damage type and a bounding box coordinate of a damage location; A conversion ratio acquisition module, configured to acquire key feature points of the blade edge in the image data, calculate the pixel distance of the key feature points on the image data according to a pre-built blade curvature model, and calculate the conversion ratio between the pixel distance and the actual size; A damage actual size acquisition module, which is configured to calculate the size of the bounding box based on the bounding box coordinates of the damage location, and then obtain the actual size of the blade damage to be detected in combination with the conversion ratio; The maintenance plan generating module is configured to compare the damage location and the actual size of the damage with the standards in the aircraft maintenance manual, and output an inspection report and a maintenance plan.

10. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement a fully automatic inspection method for high-pressure turbine blades of an aircraft engine as described in any one of claims 1-8.

Images