Infrared-assisted lightweight precision machining method and system for ducted fan housing

Abstract

This application relates to the field of precision machining and measurement technology, specifically to an infrared-assisted lightweight precision machining method and system for ducted fan housings. The method includes: acquiring three-dimensional coordinate data of the internal structure of the ducted fan housing; acquiring three-dimensional geometric data of the surface of the ducted fan housing; rigidly registering and fusing the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual housing model in a unified coordinate system; extracting wall thickness distribution data from the virtual housing model; generating differentiated machining paths for the ducted fan housing based on the wall thickness distribution data; and performing CNC precision machining according to the differentiated machining paths, combined with positioning information collected by an infrared positioning device. This application can avoid machining damage in thin-walled areas and ensure product consistency.

Description

Technical Field

[0001] This application relates to the field of precision machining and measurement technology, specifically to an infrared-assisted lightweight precision machining method and system for ducted fan housings. Background Technology

[0002] As a key precision component in aerospace and high-end equipment, the manufacturing precision of ducted fan housings directly determines the overall operational efficiency and service life of the machine. In the field of lightweight precision machining technology, for thin-walled rotating parts like ducted fan housings with complex internal flow channel structures, CNC machining technology is typically required to achieve precise material removal.

[0003] In related technologies, the machining scheme mainly plans the CNC machining tool trajectory through a pre-established theoretical three-dimensional model and generates a unified machining path based on the model for cutting. This scheme relies solely on the geometric information of the theoretical model for programming, failing to fully consider individual differences in the blank's casting errors, heat treatment deformation, and the actual distribution of its internal structure. This leads to easy interference with the thin-walled areas inside the shell during machining, posing a risk of damaging the shell. Furthermore, the lack of a unified benchmark constraint on the relative positional relationship between the external surface and the internal stops and positioning holes means that while the machined shell may meet individual dimensional requirements, key assembly relationships are difficult to guarantee, which is detrimental to ensuring product consistency and reliability. Summary of the Invention

[0004] Therefore, it is necessary to provide an infrared-assisted lightweight finishing method and system for ducted fan housings that can avoid processing damage in thin-walled areas and ensure product consistency, in order to address the above-mentioned technical problems.

[0005] In a first aspect, this application provides an infrared-assisted lightweight finishing method for ducted fan housings, the method comprising: Obtain the three-dimensional coordinate data of the internal structure of the ducted fan housing; the internal structure includes the stop and positioning holes used for positioning and installation. Obtain the surface three-dimensional geometric data of the ducted fan casing; The surface three-dimensional geometric data and the internal structure three-dimensional coordinate data are rigidly registered and fused to obtain a virtual shell model in a unified coordinate system. Extract wall thickness distribution data from the virtual shell model; Based on wall thickness distribution data, a differentiated machining path for the ducted fan housing is generated; wherein, the cutting parameters of each machining area on the differentiated machining path are set differently according to the wall thickness distribution of the corresponding machining area. Based on the positioning information collected by the infrared positioning equipment, CNC precision machining is performed according to the differentiated machining path.

[0006] Secondly, this application provides an infrared-assisted lightweight precision machining system for ducted fan housings, the system comprising: The internal structure acquisition module is used to acquire the three-dimensional coordinate data of the internal structure of the ducted fan housing; wherein, the internal structure includes a stop and positioning holes for positioning and installation; The surface scanning module is used to acquire the three-dimensional geometric data of the ducted fan casing surface; The registration and fusion module is used to rigidly register and fuse the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system. The wall thickness extraction module is used to extract wall thickness distribution data from the virtual shell model; The path generation module is used to generate differentiated machining paths for the ducted fan housing based on wall thickness distribution data; wherein, the cutting parameters of each machining area on the differentiated machining path are set differently according to the wall thickness distribution of the corresponding machining area. The machining execution module is used to perform CNC precision machining according to differentiated machining paths by combining the positioning information collected by the infrared positioning device.

[0007] The infrared-assisted lightweight finishing method and system for the ducted fan housing described above have at least the following beneficial effects: 1) The measured internal structure 3D coordinate data and the measured surface 3D geometric data are rigidly registered and fused into a virtual shell model in a unified coordinate system. This makes the geometric model used for machining programming no longer an idealized theoretical model, but a real model that includes the casting error and heat treatment deformation of the current blank. The wall thickness distribution data extracted from this model reflects the real wall thickness at various parts of the shell, thereby allocating conservative cutting parameters to thin-walled areas. This avoids interference between the tool path and the thin-walled areas inside the shell from the root, eliminating the risk of machining damage. 2) Based on the virtual shell model, a machining path with different cutting parameters as the wall thickness is generated independently for each blank to be processed. The individual differences of the blank are incorporated into the parameter design basis, eliminating the adverse effects of the traditional unified programming mode on product consistency and ensuring that each finished product meets the design wall thickness and accuracy requirements. 3) During the CNC precision machining stage, the positioning information collected in real time by the positioning equipment is combined to compensate online for the deviation between the actual and theoretical positions of the housing. This ensures that the differentiated path planned offline is accurately reproduced on the machine tool, realizing an information loop between offline planning and online execution, and ensuring that the differentiated cutting strategy is executed correctly. Attached Figure Description

[0008] Figure 1 A cross-sectional structural schematic diagram of a ducted fan housing provided in an embodiment of this application; Figure 2A flowchart of an infrared-assisted lightweight finishing method for a ducted fan housing provided in this application embodiment; Figure 3 This is a schematic diagram of an infrared-assisted lightweight finishing system for a ducted fan housing, provided as an embodiment of this application. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0010] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms; these terms are only used to distinguish one element from another. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more.

[0011] The infrared-assisted lightweight precision machining method for ducted fan housings provided in this application embodiment can be applied to precision machining scenarios for ducted fan housings. The ducted fan housing refers to a thin-walled precision component in the aerospace and high-end equipment fields, used to house ducted fans and form airflow channels. Its exterior is a complex streamlined curved surface, and its interior contains functional structures for positioning and installation. For example... Figure 1 As shown, this functional structure includes a stop and a positioning hole on the inner wall of the housing. The stop is a stepped structure on the inner wall of the housing used for positioning and mating with adjacent components (such as a motor mount); the positioning hole is a cylindrical hole with a specific diameter on the inner wall of the housing used for passing through positioning components. The stop and positioning hole serve as positioning references during housing assembly, and the accuracy of their spatial positions directly determines the assembly alignment precision.

[0012] In this application scenario, the manufacturing of ducted fan housings typically involves multiple consecutive stages, including casting, process inspection, heat treatment control, dimensional finishing, and performance verification. The finishing process is responsible for transforming the blank state into the final geometric accuracy. Since the external curved surfaces and internal functional structures of the housing are formed from different parts during the casting process, their spatial relationships are no longer in an ideal design state after heat treatment deformation. Therefore, there are often differences between the theoretical model used for machining programming and the actual blank. If machining is still performed using uniform parameters based on the theoretical model, it can easily lead to overcutting in some areas, endangering structural strength, or undercutting in some areas, increasing redundant weight. Furthermore, the lack of a unified coordinate reference between the outer surface and the internal structure makes it difficult to guarantee the critical assembly relationships between the stop and the outer surface, and between the positioning holes and the theoretical axis.

[0013] Based on this, this application provides an infrared-assisted lightweight finishing method for ducted fan housings. First, the internal structural geometry and external surface geometry of the housing are obtained using two complementary measurement methods. Then, the two types of measured data are rigidly registered to a unified coordinate system using markers set on the housing surface, thereby reconstructing a virtual housing model consistent with the actual blank. Next, the actual wall thickness distribution data is extracted from this model, and the actual wall thickness distribution drives the generation of differentiated machining paths, ensuring that the cutting parameters match the actual load-bearing capacity of various parts of the housing. Finally, during the machining execution stage, the actual pose of the housing is measured online using a positioning device, and execution deviations are compensated, allowing the offline planning results to be accurately reproduced on the machine tool. The offline stage addresses the problem of planning according to the actual blank, while the online stage addresses the problem of accurate execution according to the plan; the two constitute a complete information loop.

[0014] In some embodiments, an infrared-assisted lightweight finishing method for ducted fan housings is provided, such as... Figure 2 As shown, the method includes the following steps: Step S201: Obtain the three-dimensional coordinate data of the internal structure of the ducted fan housing.

[0015] The three-dimensional coordinate data of the internal structure refers to the three-dimensional geometric information describing the spatial position and orientation of the functional structure inside the shell. For example, it may include the axial position and orientation of the stop, the position and normal of the end face, the radius and length of the cylindrical surface, and the axial direction, diameter, and bottom position of the positioning hole. Since the stop and positioning hole are located inside the shell and cannot be directly observed from the outside, optionally, the three-dimensional coordinate data of the internal structure can be obtained through X-ray tomography. The specific acquisition process can be found in subsequent embodiments. This data is located in the first coordinate system established by the X-ray tomography scan.

[0016] Step S202: Obtain the surface three-dimensional geometric data of the ducted fan housing.

[0017] The surface three-dimensional geometric data refers to the three-dimensional geometric information describing the actual shape of the outer surface of the shell, such as surface three-dimensional point cloud data. The geometric accuracy of the outer surface of the shell determines its aerodynamic performance, and the outer surface is the main target for material removal in lightweight finishing processes; therefore, it is necessary to obtain its actual geometry with high precision. Optionally, the surface three-dimensional geometric data can be acquired through binocular structured light scanning, and the specific acquisition process can be found in subsequent embodiments. This data is located in the second coordinate system established by the optical scanning device.

[0018] Step S203: Rigidly register and fuse the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system.

[0019] Rigid registration fusion refers to the process of aligning and merging two sets of data to the same coordinate system through rotation and translation while maintaining the relative geometric relationships within the data. The virtual shell model refers to the fused digital shell model that matches the current real blank, containing complete geometric information of both the outer surface and internal structure of the shell. Since the data in step S201 is located in the first coordinate system and the data in step S202 is located in the second coordinate system, the origins and directions of the two coordinate systems are different. If they cannot be unified to the same coordinate system, the spatial correspondence between the inner and outer walls cannot be established, and the processing path cannot simultaneously avoid the internal structure and the outer surface. Optionally, at least three non-collinear marker points can be set on the shell surface. Using the coordinates of the same batch of marker points in the two coordinate systems, the rigid transformation matrix between the two coordinate systems is calculated, thereby completing the registration fusion. The specific process can be found in subsequent embodiments.

[0020] Step S204: Extract wall thickness distribution data from the virtual shell model.

[0021] Here, wall thickness distribution data refers to data that characterizes the actual wall thickness value and its spatial location at various points on the shell. Wall thickness is a direct measure of the load-bearing capacity and machining allowance of a thin-walled shell: the smaller the wall thickness at a certain point, the higher the risk of deformation and damage under cutting forces. Optionally, inner wall surface data and outer wall surface data can be separated from the virtual shell model. Rays are emitted from each vertex in the inner wall surface data in the opposite direction to its normal direction. The intersection of the ray and the outer wall surface data is taken as the corresponding point of that vertex. The distance between the vertex and the corresponding point is the wall thickness value at that point. Then, the wall thickness value is associated with the spatial location to generate wall thickness distribution data. The specific extraction process can be found in subsequent embodiments.

[0022] Step S205: Based on the wall thickness distribution data, generate a differentiated processing path for the ducted fan housing; In this differentiated machining path, the cutting parameters for each machining region are set differently according to the wall thickness distribution of the corresponding machining region. A differentiated machining path refers to a machining path where the cutting parameters are set differently depending on the machining region. From a mechanical perspective, the bending stiffness of a thin-walled structure decreases significantly as the wall thickness decreases. The smaller the wall thickness, the greater the risk of elastic deformation and chatter caused by the cutting force. Therefore, regions with smaller wall thicknesses must use more conservative cutting parameters. On the other hand, if the cutting parameters change abruptly between adjacent regions, the tool will vibrate and develop tool marks due to the sudden change in cutting force when passing through the region boundary. Optionally, the machining region can be divided into a first machining region with a larger wall thickness, a second machining region with a smaller wall thickness, and a third machining region located between the two, based on the wall thickness distribution data. A larger cutting parameter is assigned to the first machining region, a proportionally reduced cutting parameter is assigned to the second machining region, and a cutting parameter that gradually changes with position is assigned to the third machining region, thereby generating a machining path associated with the cutting parameters of each region. The specific process can be found in subsequent embodiments.

[0023] Step S206: Combine the positioning information collected by the infrared positioning device and perform CNC precision machining according to the differentiated machining path.

[0024] The positioning device refers to a measuring device capable of spatially positioning the workpiece. For example, it could be an infrared positioning device, or alternatively, a laser tracking device or a machine-contact measuring device. During machining, factors such as clamping force release, workpiece thermal deformation caused by cutting heat, and machine tool geometric errors can cause dynamic deviations between the actual and theoretical poses of the workpiece. Optionally, the infrared positioning device can be used to calculate the actual pose of the housing in the machine tool coordinate system in real time, calculate the pose deviation between the actual and theoretical poses, perform rigid body transformation compensation on the differentiated machining path according to the pose deviation, and execute CNC finishing according to the compensated machining path. This allows the offline planning results to be accurately reproduced on the machine tool. The specific compensation process can be found in subsequent embodiments.

[0025] The infrared-assisted lightweight finishing method for ducted fan housing described above acquires the three-dimensional coordinate data of the internal structure and the three-dimensional geometric data of the surface of the ducted fan housing. Through rigid registration and fusion, a virtual housing model in a unified coordinate system is generated. This ensures that the geometric model used for machining programming is no longer an idealized theoretical model, but a real model incorporating the casting errors and heat treatment deformation of the current blank. The extracted wall thickness distribution data reflects the actual wall thickness at various points on the housing, allowing for the allocation of conservative cutting parameters to thin-walled areas. This fundamentally avoids interference between the tool path and the thin-walled areas inside the housing, helping to eliminate the risk of machining damage. Furthermore, based on the wall thickness distribution data, a machining path with differentiated cutting parameters is independently generated for each blank to be machined. Incorporating the individual differences of the blanks into the parameter design basis helps ensure that each finished product meets the designed wall thickness and accuracy requirements. Finally, during the CNC finishing stage, the deviation between the actual and theoretical poses of the housing is compensated online using real-time positioning information collected by the positioning equipment. This achieves an information loop between offline planning and online execution, ensuring that the differentiated cutting strategy is correctly executed.

[0026] In some embodiments, obtaining the three-dimensional coordinate data of the internal structure of the ducted fan housing includes: performing X-ray tomography on the ducted fan housing to obtain a tomography dataset; performing three-dimensional reconstruction on the tomography dataset to obtain a three-dimensional model of the housing including the stop and the positioning hole; and extracting the three-dimensional coordinate data of the stop and the positioning hole from the three-dimensional model of the housing.

[0027] X-ray computed tomography (CT) is a non-destructive testing technique that uses X-rays to penetrate the object under test, acquire transmission projections from multiple angles, and reconstruct the object's internal three-dimensional structure. Specifically, a housing can be clamped onto the workpiece turntable of an industrial CT scanner. The turntable is controlled to rotate in preset angular increments. X-ray beams emitted from an X-ray source penetrate the housing and are received by a detector, generating projection images at various angles. Optionally, preprocessing such as defect correction, normalization, and beam hardening correction can be performed on the projection images. Then, a filtered back-projection algorithm or iterative reconstruction algorithm can be used to reconstruct the three-dimensional structure from the projection images, obtaining three-dimensional volume data. The inner and outer surfaces of the housing are then separated from the three-dimensional volume data using isosurface extraction. The inner and outer surfaces are then separated and smoothed, resulting in a three-dimensional model of the housing including the stop and positioning holes.

[0028] Finally, the three-dimensional coordinate data of the stop and the positioning hole are extracted from the three-dimensional model of the shell. Specifically, since the stop is a stepped structure and the positioning hole is a cylindrical hole structure of a specific diameter, their local curvature characteristics are significantly different from those of the smooth inner wall. Therefore, the curvature of each sampling point on the inner wall can be calculated first, and the feature areas of the stop and the positioning hole can be identified based on the curvature distribution. Then, cylindrical surface fitting is performed on each feature area, that is, the axial position, axial direction and radius that minimize the sum of the squares of the radial distance and radius deviation of each sampling point are solved. Combined with the axial range of the feature area, the three-dimensional coordinate data of the stop and the positioning hole are obtained.

[0029] In the above embodiments, the three-dimensional coordinate data of the internal structure of the shell can be directly obtained by X-ray tomography. The actual spatial position of the stop and positioning hole can be obtained completely without dissection or contact with the shell. Moreover, the data is located in the first coordinate system used for subsequent registration, which provides a benchmark for the unified integration of internal and external geometry and helps to ensure the integrity and accuracy of the internal assembly benchmark information.

[0030] In some embodiments, acquiring the surface three-dimensional geometric data of the ducted fan housing includes: projecting an coded speckle pattern onto the surface to be scanned of the ducted fan housing using an optical scanning device, and acquiring an image of the speckle pattern reflected from the surface to be scanned; performing stereo matching on the acquired speckle pattern image to obtain a disparity map of the surface to be scanned; and calculating the three-dimensional point cloud data of the surface to be scanned based on the disparity map and the calibration parameters of the optical scanning device, as the surface three-dimensional geometric data of the ducted fan housing.

[0031] Binocular structured light scanning is an optical measurement technique that projects a coded pattern onto the surface being measured, which is then simultaneously acquired by two cameras and the three-dimensional morphology of the surface is reconstructed based on the principle of triangulation. The random texture of the coded speckle pattern provides distinguishable features for stereo matching. Specifically, the intrinsic and extrinsic parameters of the optical scanning device need to be calibrated before scanning: the intrinsic parameters include the equivalent focal length. , and principal point coordinates The external parameters include the relative positions of the optical centers of the two cameras, and the distance between the two optical centers forms the baseline. For example, calibration can be performed based on Zhang's calibration method. After calibration, an coded speckle pattern is projected onto the surface to be scanned, and the reflected images are simultaneously acquired by the left and right cameras. Optionally, for highly reflective metal surfaces, a removable developer can be pre-sprayed to improve surface diffuse reflection.

[0032] After acquiring the left and right images, stereo matching is performed between them. This involves finding the corresponding pixel on the same surface in the right image for each pixel in the left image. The difference in the x-coordinates of the two pixels is the disparity. The disparities of all pixels constitute a disparity map. Then, a 3D point cloud is calculated based on the disparity map and calibration parameters: disparity is inversely proportional to depth.

[0033] in, For pixels The disparity value at the location is obtained; after obtaining the depth, the spatial coordinates of the point are solved by inverse solving based on the pinhole camera model.

[0034] By storing the spatial points corresponding to all pixels of the surface to be scanned in sequence, the three-dimensional point cloud data of the surface can be obtained.

[0035] In some embodiments, the optical scanning device can be controlled to move along a preset trajectory to repeatedly perform the above-described projection, acquisition, stereo matching, and calculation on multiple surfaces to be scanned on the ducted fan housing, obtaining three-dimensional point cloud data for each surface to be scanned. The three-dimensional point cloud data are then stitched and fused to obtain the surface three-dimensional geometric data of the ducted fan housing. The preset trajectory is pre-planned, and a 20% to 30% overlap area is maintained between adjacent scanning positions. For example, the stitching can be performed by calculating a rigid transformation matrix between adjacent point clouds using at least three non-collinear marker points set on the housing surface, and then sequentially connecting each local point cloud to a unified coordinate system. Optionally, global optimization can be performed after stitching to eliminate accumulated errors. This surface three-dimensional geometric data is located in a second coordinate system, which is the coordinate system established by the optical scanning device.

[0036] In the above embodiments, the three-dimensional geometric data of the shell surface is obtained by binocular structured light scanning and multi-view stitching. Complete three-dimensional information from one viewpoint can be obtained in one shot. The measurement process is non-contact and highly efficient, which is conducive to accurately depicting the true shape of the outer surface of the shell and providing high-precision external geometric input for subsequent registration fusion and wall thickness extraction.

[0037] In some embodiments, rigid registration and fusion of surface three-dimensional geometric data and internal structural three-dimensional coordinate data are performed to obtain a virtual shell model in a unified coordinate system. This includes: setting at least three non-collinear marker points on the surface of the ducted fan shell; acquiring the first coordinates of each marker point in a first coordinate system; wherein the first coordinate system is the coordinate system established when acquiring the internal structural three-dimensional coordinate data; acquiring the second coordinates of each marker point in a second coordinate system; wherein the second coordinate system is the coordinate system established when acquiring the surface three-dimensional geometric data; calculating a rigid transformation matrix from the second coordinate system to the first coordinate system based on the first and second coordinates of each marker point; transforming the surface three-dimensional geometric data to the first coordinate system using the rigid transformation matrix and fusing it with the internal structural three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system.

[0038] In this context, marker points refer to physical targets set on the shell surface that can be clearly identified and accurately centered in both X-ray tomography and optical scanning. These can be high-contrast targets such as spherical targets or arrays of planar dots. Since the marker points are a group of physical points on the shell surface, the difference between their two sets of coordinates in the two coordinate systems precisely reflects the rigid relationship between the two coordinate systems; therefore, the marker points act as a bridge connecting the two coordinate systems. Three non-collinear marker points represent the minimum constraints for determining the rigid transformation in three-dimensional space. Optionally, the marker points can be set in a flat or nearly flat area on the outer surface of the shell, and should be as uniformly distributed, non-collinear, and not nearly coplanar as possible to improve the stability of the solution. For example, there can be four marker points, arranged at both ends and both sides of the outer surface of the shell.

[0039] Specifically, when performing the above-mentioned X-ray tomography scan, the first coordinates of each marker point in the first coordinate system can be extracted simultaneously during the reconstruction of the tomography data. During the optical scanning process described above, the second coordinates of each marker point in the second coordinate system can be extracted simultaneously during the point cloud data processing. Where the number of markers is ; ( ), Mark the point number, superscript , These represent the first and second coordinate systems, respectively, with the coordinate unit being mm. Furthermore, based on the spatial distribution order of the marker points, their relative positions, or shape recognition algorithms, a one-to-one correspondence is established between the two sets of coordinates.

[0040] After obtaining two sets of corresponding coordinates, the rigid transformation matrix from the second coordinate system to the first coordinate system can be calculated. The specific calculation process can be found in the following embodiment. Furthermore, the surface three-dimensional geometric data... Each surface point in the system is transformed to the first coordinate system using a rigid transformation matrix, i.e.

[0041] in, The total number of points in the surface point cloud. Let be a rotation matrix. It can be a translation vector; homogeneous coordinates can also be used.

[0042] in For the reason and Composition Homogeneous transformation matrix. Then, the transformed surface 3D geometric data... With internal structural three-dimensional coordinate data By merging, a virtual shell model is obtained. ;in, This represents the total number of internal feature points. This is a set union operation. Optionally, the merged virtual shell model can also be subjected to consistency checks and smoothing processing for subsequent wall thickness extraction and CNC machining programming.

[0043] In some embodiments, calculating the rigid transformation matrix based on the first and second coordinates of each marker point may include: calculating the centroid of the first coordinate and the centroid of the second coordinate of each marker point respectively to obtain first centroid coordinates and second centroid coordinates; subtracting the first centroid coordinate from each first coordinate to obtain first centroid coordinates; subtracting the second centroid coordinate from each second coordinate to obtain second centroid coordinates; calculating a rotation matrix based on each first centroid coordinate and each second centroid coordinate; calculating a translation vector based on the first centroid coordinate, the second centroid coordinate, and the rotation matrix; and combining the rotation matrix and the translation vector to obtain the rigid transformation matrix from the second coordinate system to the first coordinate system.

[0044] In this process, rigid transformation includes two components: rotation and translation. The centroid refers to the geometric center of the point set. The purpose of centroid removal is to decouple rotation and translation: after translating the origins of the two sets of coordinates to their respective centroids, the translation component is eliminated, and the remaining difference is caused only by rotation. The rotation matrix can then be solved independently on the centroid-removed coordinates.

[0045] Specifically, the coordinates of the first centroid are

[0046] The coordinates of the second centroid are:

[0047] Centroid coordinates are and Optionally, the rotation matrix can be calculated using singular value decomposition: calculate the cross-covariance matrix.

[0048] right Singular value decomposition yields The rotation matrix that minimizes the sum of the squared deviations of the centroid coordinates of each marked point is:

[0049] in, , It is an orthogonal matrix. It is a singular value diagonal matrix.

[0050] It should be noted that when the determinant of the orthogonal matrix is ​​+1, it corresponds to a pure rotation; when it is -1, it corresponds to a combination of rotation and reflection. However, physically, only rotation and translation relationships are possible between two coordinate systems. Therefore, optionally, a reflection check and correction can be performed on the rotation matrix: if... Then, the singular vector corresponding to the minimum singular value is inverted for correction, i.e.

[0051] Make the corrected matrix correspond to a pure rotation.

[0052] Once the rotation matrix is ​​determined, the translation vector can be directly determined by the relationship between the centroids of the two coordinate systems. Specifically, the mathematical form of a rigid transformation is: Since the centroid is the geometric center of the point set, a rigid transformation will necessarily map the centroid of the second coordinate system to the centroid of the first coordinate system, i.e.

[0053] By rearranging the terms, we can obtain the result.

[0054] Finally, combine the rotation matrices. Translation vector This yields the rigid transformation matrix; for example, it can be represented in homogeneous coordinates as follows: matrix

[0055] In the above embodiments, by using the same batch of markers set on the shell surface, rotation and translation are decoupled by centroid removal, and singular value decomposition is used for closed-form solution and correction by reflection verification. The entire solution process does not require iteration, is optimal in the least squares sense, and can ensure that the obtained transformation is a physically legal pure rigid transformation. This unifies the external surface geometry and internal assembly datum to the same coordinate system, which is beneficial to ensure the geometric consistency of subsequent wall thickness extraction and machining programming.

[0056] In some embodiments, extracting wall thickness distribution data from a virtual shell model includes: separating inner wall surface data and outer wall surface data from the virtual shell model; emitting a ray in the opposite direction to the normal direction of each vertex in the inner wall surface data; determining the intersection points of each ray with the outer wall surface data and marking the intersection points as the corresponding points of the corresponding vertex on the outer wall surface data; calculating the distance between each vertex and its corresponding point to obtain the wall thickness value at each corresponding point pair; and generating wall thickness distribution data based on the wall thickness value and spatial location at each corresponding point pair.

[0057] The wall thickness extraction method based on ray projection measures the thickness along the normal direction of the wall surface, which is consistent with the physical definition of wall thickness. It should be noted that the normal direction is defined as follows: in this embodiment, the normal direction of each vertex in the inner wall surface data is uniformly directed towards the interior of the virtual shell model, that is, towards the cavity side inside the shell; correspondingly, rays emitted in the opposite direction of the normal direction pass through the shell wall from the inner wall surface and point towards the outer wall surface, consistent with the direction of wall thickness measurement.

[0058] Specifically, the virtual shell model can be first separated into inner wall surface data and outer wall surface data. Optionally, separation can be based on the normal direction: the normal direction of each point on the outer surface of the shell points outwards from the model, while the normal direction of each point on the inner wall points inwards from the model, thus determining the location of each point; alternatively, the location can be determined based on the spatial coordinates of the points (such as radius and axial position). The separation result is denoted as the inner wall surface data set. and outer wall surface data set .

[0059] Furthermore, the normal direction of each vertex in the inner wall surface data is calculated, and its orientation is unified: for point cloud data, multiple nearest neighbors of the vertex to be determined can be selected for plane fitting, and the normal vector of the fitted plane can be used as the normal direction of the vertex; for triangular mesh data, the weighted average of the normals of its adjacent faces can be taken as its normal direction. An exemplary method for unifying the orientation can be: calculating the average centroid position of the inner wall surface data. For any point in the inner wall data and its normal direction ,like This indicates Point inwards from the model, maintain that direction, otherwise... Invert the normals. After this process, the normals of all inner wall vertices will point inwards to the model.

[0060] Once the normal direction is determined, a ray can be emitted from each vertex in the inner wall data in the opposite direction to the normal direction of that vertex, and the intersection point of the ray with the outer wall data can be determined. Specifically, let the ray in the inner wall data be the _th_ vertex. vertices The normal vector is (Unit vector, pointing into the model), then the parametric equation of the ray is:

[0061] in, Let be the distance parameter along the ray direction. The negative sign indicates that the ray propagates in the opposite direction to the normal, that is, from the vertex of the inner wall, through the shell wall, and towards the outer wall. Optionally, the intersection point can be calculated using the method of finding the intersection of the ray and the triangular facet: represent the outer wall data as a triangular mesh, and let the unit normal vector of the triangular facet be... Let the direction of the ray be . Then, assuming the ray direction is not parallel to the triangle plane, the intersection point parameter is...

[0062] The coordinates of the intersection point are ,in Let be a vertex of a triangular facet; then determine whether the intersection point is inside the triangular facet. An intersection point inside the triangular facet is a valid intersection point.

[0063] After obtaining the valid intersection point, the vertex of the inner wall surface Its corresponding point on the outer wall surface (i.e., valid intersection point) That is, forming a pair of corresponding points, the wall thickness value at that pair of points. Euclidean distance between two points

[0064] The unit is mm; since the ray direction is a unit vector, the wall thickness is numerically equal to the ray parameter value at the effective intersection point, i.e. Finally, by associating the wall thickness values ​​at each valid corresponding point pair with their spatial locations, wall thickness distribution data can be generated.

[0065] in, The number of valid corresponding point pairs.

[0066] In the above embodiments, a ray is emitted in the opposite direction to the normal of the vertex of the inner wall surface, and the corresponding point is determined by the intersection of the ray with the outer wall surface. The wall thickness is measured along the normal direction of the wall surface, which is consistent with the physical definition of the wall thickness. The obtained wall thickness distribution data truly reflects the wall thickness and its spatial position at various points in the current blank, providing an accurate basis for the generation of differentiated processing paths and helping to improve the accuracy and efficiency of wall thickness measurement.

[0067] In some embodiments, determining the intersection points of each ray with the outer wall data includes: when there are multiple intersection points between the ray and the outer wall data, determining the intersection point closest to the vertex as the corresponding point; when there are no intersection points between the ray and the outer wall data, removing the corresponding vertex, and the removed vertex is not included in the calculation of the wall thickness value.

[0068] A ray exiting from the vertex of the inner wall may pass through multiple wall surfaces. For example, when the ray path passes through a stop step or reinforcing rib, it will intersect with the outer wall data at multiple points. In this case, the intersection point closest to the vertex of the inner wall corresponds to the wall thickness at that location, while the more distant intersection points correspond to other wall structures. Therefore, one of the intersection points can be selected. and The smallest intersection point is considered a valid intersection point. To ensure the intersection point is located in the positive direction of the ray, the minimum value is taken to ensure that the corresponding point is the outer wall point of the local wall surface. On the other hand, when the vertex of the inner wall surface is located at the edge of the shell opening or in a data missing area, the ray may not intersect with the outer wall surface data. Such vertices cannot obtain valid wall thickness values, so they can be discarded and marked in the wall thickness distribution data to avoid invalid data contaminating the wall thickness distribution.

[0069] In the above embodiments, by taking the nearest intersection point from multiple intersection points and removing vertices without intersection points, it is ensured that each wall thickness value corresponds to the true and nearest inner and outer wall point pair, avoiding erroneous wall thicknesses at complex structures such as steps and stiffeners, as well as at the edges of openings, which helps to improve the reliability of wall thickness distribution data in complex structural regions.

[0070] In some embodiments, a differentiated machining path for the ducted fan housing is generated based on wall thickness distribution data, including: obtaining the design wall thickness value of the ducted fan housing; multiplying the design wall thickness value by a preset proportional coefficient to obtain a wall thickness threshold; determining regions with wall thickness values ​​not less than the wall thickness threshold as thick-wall candidate regions and regions with wall thickness values ​​less than the wall thickness threshold as thin-wall candidate regions; dividing the strip-shaped regions on both sides of the boundary line between the thick-wall candidate regions and the thin-wall candidate regions, with a distance from the boundary line not exceeding a preset transition width, into a third machining region; determining the regions in the thick-wall candidate regions other than the third machining region as a first machining region and the regions in the thin-wall candidate regions other than the third machining region as a second machining region; and then assigning differentiated cutting parameters to the first machining region, the second machining region, and the third machining region, and generating a differentiated machining path associated with the cutting parameters of each machining region.

[0071] The design wall thickness value refers to the nominal wall thickness taken from the shell design model or design drawings, denoted as . The unit is mm. Multiply the designed wall thickness value by the preset scaling factor. The wall thickness threshold can then be obtained. ;in, It is a dimensionless proportionality coefficient, determined based on machining experience and tool parameters, with a typical value range of 0.8 to 1.2.

[0072] Specifically, all valid points can be divided into two categories of candidate regions based on a wall thickness threshold: thick-walled candidate regions. and thin-walled candidate regions The boundary between the two types of candidate regions This represents the boundary line between thick and thin-walled candidate regions on the meshed surface of the shell. Furthermore, to achieve a continuous transition of cutting parameters, a strip-shaped transition region can be reserved on both sides of the boundary line. For example, let the transition width be... Its typical value can be 0.5 to 2 times the tool diameter; the strip-shaped area on both sides of the dividing line, with the shortest distance from the dividing line not exceeding the transition width, is divided into the third machining area.

[0073] in, For point To the dividing line The shortest distance. Subsequently, the area in the thick-walled candidate region, excluding the third processing region, is designated as the first processing region. The area outside the third processing area in the thin-walled candidate region is designated as the second processing area. ,in" "Represents the set difference operation. When the thick-walled candidate region or the thin-walled candidate region is empty (for example, when the wall thickness of the blank is consistently greater than or less than the wall thickness threshold), the thick-walled candidate region and the thin-walled candidate region are re-divided by the median wall thickness of all valid points, so that the first processing region and the second processing region are not empty.

[0074] It should be noted that, according to this division, the first processing area, the second processing area, and the third processing area do not overlap with each other and together cover all valid points; furthermore, any point in the third processing area does not belong to the first processing area or the second processing area, and its distance to both areas is greater than zero, so that the interpolation coefficients take continuous values ​​within the transition area, avoiding the interpolation from degenerating to the values ​​of the two endpoints.

[0075] In the above embodiments, by first dividing the candidate regions of thick and thin according to the wall thickness threshold, and then independently dividing the strip-shaped regions on both sides of the dividing line into the third processing region, the three processing regions do not overlap and completely cover all effective points, providing a geometric basis for the continuous and gradual allocation of cutting parameters within the transition region, which helps to avoid parameter jumps caused by overlapping region affiliations.

[0076] In some embodiments, assigning differentiated cutting parameters to a first machining area and a second machining area includes: determining a first cutting parameter combination for the first machining area; the first cutting parameter combination includes at least a first depth of cut and a first feed rate; multiplying the first depth of cut by a first preset coefficient to obtain a second depth of cut for the second machining area; multiplying the first feed rate by a second preset coefficient to obtain a second feed rate for the second machining area; wherein both the first preset coefficient and the second preset coefficient are less than 1; and combining the second depth of cut and the second feed rate to obtain a second cutting parameter combination.

[0077] In this context, the cutting parameter combination refers to a set of process parameters that work together to determine the cutting process. Specifically, a first cutting parameter combination can be determined for the first machining area. ;in, The first depth of cut is measured in mm and is determined based on the machining allowance and tool parameters. This is the first feed rate, in mm / rev, determined based on surface quality requirements and tool parameters. The first spindle speed, in rpm, is determined based on the allowable cutting speed and tool diameter. The first machining area has a large wall thickness and high structural rigidity, allowing for larger cutting parameters to ensure machining efficiency. It should be noted that the cutting speed... With spindle speed They are not two independent physical quantities; they are related by the tool diameter. Fixed connection, that is

[0078] Among them, cutting speed The unit is m / min, and the tool diameter is... The unit is mm, spindle speed The unit is rpm; therefore, in this embodiment, each cutting parameter combination uses depth of cut, feed rate, and spindle speed as independent settings. The cutting speed is calculated from the tool diameter and spindle speed according to the above formula, and is only used as a reference for process verification and machine tool status monitoring, thereby ensuring that the parameter combination is physically self-consistent.

[0079] Furthermore, a second combination of cutting parameters can be determined for the second machining area. The second machining area has a smaller wall thickness and lower structural stiffness. The bending stiffness of a thin-walled structure decreases significantly with decreasing wall thickness; the bending stiffness of the plate is approximately proportional to the cube of the wall thickness. Therefore, the cutting force should be reduced accordingly to match the load-bearing capacity of this area. Since the cutting force increases with increasing depth of cut and feed rate, proportionally reducing these two parameters will reduce the cutting force accordingly. For example, the second depth of cut can be... The second feed rate can be ;in, The first preset coefficient is determined based on the ratio of the average wall thickness of the second processing area to the designed wall thickness. The smaller the average wall thickness, the better. The smaller the value, the more typical the range is from 0.3 to 0.7; This is the second preset coefficient, with a typical value range of 0.5 to 0.9; , All values ​​are less than 1 to ensure that the unit cutting load in the second machining area is lower than that in the first machining area. Optionally, the second spindle speed can also be reduced proportionally, i.e. ,in The third preset coefficient typically ranges from 0.6 to 1.0, in order to reduce the cutting speed and weaken the cutting heat input and chatter excitation in thin-walled areas.

[0080] In some embodiments, a third cutting parameter combination can also be determined for the third processing area: interpolation coefficients are determined based on the distances from points in the third processing area to the first processing area and to the second processing area, and the first and second cutting parameter combinations are linearly interpolated according to the interpolation coefficients to obtain a third cutting parameter combination for the third processing area; wherein, the closer the point is to the first processing area, the closer the third cutting parameter combination is to the first cutting parameter combination.

[0081] Specifically, let's set For any point within the third processing region, interpolation coefficients can be defined.

[0082] in, For point The shortest distance to the first processing area For point The shortest distance to the second processing area. Since the third processing area does not overlap with the first and second processing areas, the distance from any point within the transition zone to either area is greater than zero. exist Continuous values ​​within an open interval: when point When near the boundary of the first processing area, Approaching 0; when point When near the boundary of the second processing area, It approaches 1. That is to say, This indicates the relative degree to which points within the transition zone approach thick-walled or thin-walled regions. Third cutting parameter combination. That is, it is obtained by linear interpolation using interpolation coefficients:

[0083]

[0084]

[0085] From the interpolation formula, we know that: Approaching 0 Approaching , Approaching 1 Approaching Therefore, the cutting parameters change continuously along the transition zone from the thick-walled region to the thin-walled region, and there are no abrupt changes in parameters.

[0086] Finally, differentiated machining paths associated with the cutting parameters of each machining area can be generated: based on the geometric characteristics of each machining area, machining paths for each machining area are generated using methods such as line cutting, circular cutting, or helical cutting. The corresponding cutting parameters are then combined and associated with each machining path, and finally spliced ​​together to form a complete differentiated machining path. Each machining path contains a sequence of trajectory points and their corresponding cutting parameters. At the junction of trajectories in adjacent areas, since the cutting parameters of the machining path in the third machining area change continuously and gradually, the cutting load transitions smoothly when the tool crosses the boundary of the area, without causing vibrations caused by abrupt parameter changes. Optionally, the differentiated machining path can be transformed from the first coordinate system where the virtual shell model is located to the machine tool coordinate system: After the shell is clamped on the machine tool, for example, the coordinates of each marked point on the shell surface in the machine tool coordinate system can be measured by an infrared positioning device or a machine contact probe, and rigidly registered with the coordinates of the same batch of marked points in the virtual shell model (the registration algorithm is the same as the calculation process of the rigid transformation matrix mentioned above), to obtain the rigid transformation matrix from the first coordinate system to the machine tool coordinate system, and the differentiated machining path is transformed to the machine tool coordinate system through this matrix, so that the offline planned path and the actual workpiece on the machine tool are docked at the coordinate level.

[0087] In the above embodiments, larger cutting parameters are used in the thick-walled region to ensure efficiency, while the cutting parameters in the thin-walled region are reduced by a coefficient of less than 1 to match its lower stiffness. The parameters in the transition region are continuously and gradually changed with position according to the interpolation coefficient, so that the cutting parameters are accurately matched with the actual wall thickness at various parts of the shell and smoothly connected. This is beneficial to ensure the processing safety of the thin-walled region while taking into account both processing efficiency and surface quality.

[0088] In some embodiments, the process of performing CNC finishing according to a differentiated machining path further includes: establishing a mapping relationship between thermal diffusion characteristic time and wall thickness in advance through calibration; collecting transient temperature data of the machined surface after the tool passes over it; extracting the thermal diffusion characteristic time of the transient temperature data; inverting the remaining wall thickness at the machined surface based on the thermal diffusion characteristic time and mapping relationship; and correcting the cutting parameters of the corresponding machining area based on the inverted remaining wall thickness.

[0089] The aforementioned wall thickness distribution data is static data acquired offline before machining. However, during finishing, material is gradually removed along the tool path, and the wall thickness at various points on the shell is reduced in real time. Furthermore, local areas of the blank may exhibit casting deviations inconsistent with the virtual shell model. If differentiated cutting is performed solely based on the static wall thickness distribution data, there will be a lag between the parameter allocation and the actual wall thickness during machining. Infrared thermal imaging equipment refers to devices that collect infrared radiation from an object's surface and generate a surface temperature distribution image; for example, it can be a cooled or uncooled infrared thermal imager. It should be noted that infrared thermal imaging equipment differs in function from the aforementioned infrared positioning equipment used for pose calculation, and their operating bands are also different; therefore, they are set up independently.

[0090] The physical principle is as follows: the cutting process injects heat into the machined surface, and the injected heat diffuses along the wall thickness from the machined surface to the inner wall of the shell. Since the inner cavity of the shell is an approximately adiabatic air boundary, the thinner the wall, the shorter the time required for heat diffusion to reach the inner wall boundary and cause a change in the shape of the surface temperature decay curve. Therefore, the thermal diffusion characteristic time of the transient temperature data of the machined surface after the tool passes is monotonically correlated with the remaining wall thickness at that point. In other words, the cutting process itself injects a thermal excitation into the workpiece, without the need to stop the machine or add an external excitation device. The thermal diffusion characteristic time refers to the characteristic moment when the temperature decay curve deviates from the decay law of a semi-infinite medium. Optionally, the collected temperature decay sequence can be compared with the theoretical decay curve of a semi-infinite medium under the same injected heat, and the moment when the two begin to deviate can be extracted as the thermal diffusion characteristic time. The remaining wall thickness refers to the material thickness from the machined surface to the inner wall of the shell.

[0091] Specifically, cutting thermal imaging calibration can be performed in advance on a stepped calibration test block made of the same material as the shell: the stepped calibration test block has multiple stepped surfaces of known thickness. Optionally, the stepped surfaces can be surfaces with curvature similar to the area to be machined in the shell. Each stepped surface is machined with the same cutting parameters as the actual machining, and transient temperature data after the tool passes over is collected. The thermal diffusion characteristic time corresponding to each stepped surface is extracted, and a mapping relationship between the thermal diffusion characteristic time and the known wall thickness is established. Since the thermal diffusion characteristic time is the time-domain morphological feature of the decay curve, it does not change with the amount of injected heat. Therefore, this mapping relationship is still applicable after the cutting parameters are adjusted. Moreover, since the thermal excitation method and surface state of the calibration are consistent with those of the actual machining, this mapping relationship naturally takes into account the influence of factors such as the lateral diffusion of hot spots and surface emissivity. During actual machining, the infrared thermal imaging device moves with the cutting tool and is triggered synchronously according to the tool position. It collects the temperature decay sequence of each machining surface point after the tool passes over it, extracts the thermal diffusion characteristic time, and queries the mapping relationship to obtain the remaining wall thickness at that point. The position of each machining surface point in the virtual shell model can be determined by the coordinates of the tool trajectory points recorded during synchronous triggering, through the aforementioned coordinate transformation. Then, the retrieved remaining wall thickness is compared with the wall thickness value at the corresponding position in the wall thickness distribution data. When the deviation between the retrieved remaining wall thickness and the corresponding wall thickness value exceeds a preset threshold (e.g., 5% to 10% of the corresponding wall thickness value), the wall thickness distribution data is updated according to the retrieved remaining wall thickness, and the cutting parameters of the corresponding machining area are re-determined according to the updated wall thickness. For example, the depth of cut and feed rate are proportionally corrected according to the parameter determination method of the second machining area. The acquisition frame rate of the infrared thermal imaging device can be matched with the thermal diffusivity of the shell material: the greater the thermal diffusivity of the material, the shorter the thermal diffusion characteristic time, and the higher the required acquisition frame rate. Optionally, an air curtain can be set in front of the tool to remove chips and stabilize the surface reflection background.

[0092] In the above embodiments, the heat injected into the workpiece during the cutting process is used as the measurement excitation. The infrared thermal imaging equipment inverts the true remaining wall thickness during processing and corrects the cutting parameters accordingly. The cutting parameters can track the true changes in wall thickness without stopping the machine or adding an external excitation device. This helps to eliminate the lag between static wall thickness data and the true wall thickness during processing, and further reduces the risk of processing damage in thin-walled areas.

[0093] In some embodiments, after correcting the cutting parameters of the corresponding processing area based on the inverted remaining wall thickness, the method further includes: verifying whether the heat injected per unit area corresponding to the corrected cutting parameters is within a preset invertible range; if it is not within the invertible range, then redetermining the corrected cutting parameters within the set of cutting parameters whose heat injected per unit area is within the invertible range.

[0094] The corrected cutting parameters also determine the heat injected into the workpiece in the next cutting step: if the injected heat per unit area is too small, the surface temperature rise will be lower than the resolvable level of the infrared thermal imaging equipment, and the subsequent inversion will lose its signal basis; if the injected heat per unit area is too large, the surface temperature will exceed the equipment's range or cause surface overheating, and the inversion will also fail. In other words, the correction of cutting parameters cannot only meet processing requirements, but also ensure the continued feasibility of in-situ inversion. Optionally, the lower limit of the invertible range can be determined by the minimum resolvable temperature rise of the infrared thermal imaging equipment at a given acquisition frame rate and surface emissivity, and the upper limit can be determined by the equipment's saturation threshold and the allowable temperature rise of the shell material surface; the injected heat per unit area increases with the increase of the depth of cut, for example, it can be calculated as follows:

[0095] An estimate is made, in which, Injecting heat per unit area The proportion of cutting heat flowing into the workpiece. Energy removed per unit volume of material For the back of the knife; and It can be determined through cutting tests or by consulting cutting data manuals.

[0096] Specifically, if the verification is not satisfied, the correction amount can be re-determined within the set of cutting parameters that satisfy the invertible range: for example, the depth of cut should be reduced in layers and the feed rate should be adjusted accordingly to reduce the heat injected per unit area back to the invertible range; if the set of cutting parameters that satisfy the invertible range is empty, the original correction parameters should be retained for machining, and the inversion result of the machining area should be marked as low confidence, and the inversion result of the area should not participate in parameter correction; optionally, for areas with complex internal structures such as stops and positioning holes, their inversion results can also be marked as low confidence.

[0097] In the above embodiments, the correction of cutting parameters is simultaneously constrained by the invertible range, so that the cutting parameters serve both material removal and ensure the signal quality of subsequent wall thickness inversion. Measurement and processing form a mutually reinforcing closed loop, which is conducive to ensuring the continuous reliability of in-situ wall thickness inversion throughout the entire processing.

[0098] In some embodiments, the positioning device is an infrared positioning device. An infrared positioning device refers to a device that uses infrared light for non-contact spatial positioning of a target, such as an infrared camera and infrared markers that move with the workpiece. In some embodiments, combining the positioning information collected by the positioning device, CNC finishing is performed according to a differentiated machining path, including: setting at least three non-collinear infrared markers on the ducted fan housing or the fixture holding the ducted fan housing; acquiring the coordinates of each infrared marker in the machine tool coordinate system in real time using the infrared positioning device, calculating the actual pose of the ducted fan housing in the machine tool coordinate system; calculating the pose deviation between the actual pose and the theoretical pose, and performing rigid body transformation compensation on the differentiated machining path according to the pose deviation; and performing CNC finishing according to the compensated machining path.

[0099] Infrared markers refer to actively emitting infrared light points or passively reflecting infrared light spheres, fixed to the workpiece or fixture and moving synchronously with the workpiece; infrared cameras are fixed to stationary parts of the machine tool. Pose refers to the overall position and orientation of a rigid body in space. Since three non-collinear markers can uniquely determine the spatial pose of a rigid body, the infrared camera can calculate the actual pose of the workpiece in the machine tool coordinate system by tracking the spatial positions of each infrared marker.

[0100] Specifically, before and during machining, the infrared positioning device can collect the coordinates of each infrared marker in the machine tool coordinate system in real time, and calculate the actual pose of the housing based on the marker coordinates. Then, it calculates the pose deviation between the actual pose and the theoretical pose. For example, let the actual position coordinate vector of the workpiece reference point in the machine tool coordinate system in the actual pose be denoted as... The nominal position coordinate vector of the workpiece reference point in the theoretical pose is: The positional deviation is

[0101] The theoretical pose is determined by the aforementioned coordinate transformation. Simultaneously, the rotational deviation of the workpiece's actual pose relative to its theoretical pose can be calculated from the correspondence between the actual and theoretical coordinates of each infrared marker. The solution method is the same as that for solving the rotation matrix in the rigid transformation matrix described above; position deviation With rotational deviation Together they constitute the pose deviation.

[0102] Subsequently, rigid body transformation compensation is performed on the differentiated machining paths according to the pose deviation. Specifically, the workpiece reference point can be used as the rotation center, and for each trajectory point in the path, the rotation deviation is first calculated. Rotate, then press the position deviation Translation yields the compensated machining path. For the case where only translational deviation exists, i.e. When the matrix is ​​an identity matrix, compensation can be simplified to superimposing the positional deviation onto each trajectory point of the differentiated machining path. The physical meaning of this compensation is: the tool path shifts by the same amount as the actual workpiece pose deviates from the theoretical pose, ensuring that the geometric relationship between the tool and the workpiece remains unchanged, and the machining result matches the planned intention. Figure 1 Then, the CNC system performs finishing according to the compensated machining path. Optionally, during the machining process, the infrared positioning device continuously acquires the workpiece pose in real time; if the detected position deviation exceeds the set threshold, the pose deviation is recalculated and the compensation amount is updated so that the compensation always tracks the actual state of the workpiece; wherein, the set threshold can be, for example, 0.1mm, which matches the measurement accuracy of the infrared optical tracking system.

[0103] After all trajectories have been executed, the finished ducted fan housing is output. At this point, the outer surface of the housing undergoes material removal according to a differentiated strategy: redundant material is removed from thick-walled areas to achieve weight reduction; thin-walled areas are machined with conservative parameters to maintain wall thickness and structural strength; and the relative positional relationship between the stops, positioning holes, and the outer surface is guaranteed by a unified coordinate system constraint.

[0104] In the above embodiments, the actual pose of the workpiece is calculated in real time using at least three non-collinear infrared markers, and rigid body transformation compensation is performed on the differentiated machining path according to the pose deviation. This enables the differentiated path planned offline to be accurately reproduced on the machine tool, avoiding the impact of dynamic deviations such as clamping errors and thermal deformation on machining accuracy. This is beneficial for achieving an information closed loop between offline planning and online execution.

[0105] The above text combined Figure 2 The infrared-assisted lightweight finishing method for ducted fan housing provided in this application embodiment is described in detail below. Figure 3 The system provided in the embodiments of this application will be described.

[0106] Based on the same inventive concept, this application also provides an infrared-assisted lightweight finishing system for ducted fan housings, used to implement the infrared-assisted lightweight finishing method for ducted fan housings involved in the above-described method embodiments. The solution provided by this system is similar to the solution described in the above-described method. Therefore, the specific limitations of one or more embodiments of the infrared-assisted lightweight finishing system for ducted fan housings provided below can be found in the limitations of the infrared-assisted lightweight finishing method for ducted fan housings described above, and will not be repeated here.

[0107] An infrared-assisted lightweight precision machining system for ducted fan housings, the system comprising: The internal structure acquisition module 301 is used to acquire the three-dimensional coordinate data of the internal structure of the ducted fan housing; wherein, the internal structure includes a stop and positioning holes for positioning and installation. Surface scanning module 302 is used to acquire three-dimensional geometric data of the surface of the ducted fan housing; The registration and fusion module 303 is used to rigidly register and fuse the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system. The wall thickness extraction module 304 is used to extract wall thickness distribution data from the virtual shell model; The path generation module 305 is used to generate differentiated machining paths for the duct fan housing based on wall thickness distribution data; wherein, the cutting parameters of each machining area on the differentiated machining path are set differently according to the wall thickness distribution of the corresponding machining area. The machining execution module 306 is used to perform CNC precision machining according to the differentiated machining path by combining the positioning information collected by the infrared positioning device.

[0108] The modules in the infrared-assisted lightweight precision machining system for the aforementioned ducted fan housing can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0109] In one embodiment, a computer device is provided, which may be a server.

[0110] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above method embodiments.

[0111] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in any of the above method embodiments.

[0112] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0113] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0114] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0115] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An infrared-assisted lightweight finishing method for ducted fan housings, characterized in that, The method includes: Obtain three-dimensional coordinate data of the internal structure of the ducted fan housing; wherein, the internal structure includes a stop and positioning holes for positioning and installation; Obtain the three-dimensional geometric data of the surface of the ducted fan housing; The surface three-dimensional geometric data and the internal structure three-dimensional coordinate data are rigidly registered and fused to obtain a virtual shell model in a unified coordinate system. Extract wall thickness distribution data from the virtual shell model; Based on the wall thickness distribution data, a differentiated machining path for the ducted fan housing is generated; wherein, the cutting parameters of each machining area on the differentiated machining path are set differently according to the wall thickness distribution of the corresponding machining area. Based on the positioning information collected by the infrared positioning device, CNC precision machining is performed according to the differentiated machining path.

2. The method according to claim 1, characterized in that, The acquisition of the three-dimensional coordinate data of the internal structure of the ducted fan housing includes: X-ray tomography was performed on the ducted fan casing to obtain a tomography dataset; The tomographic scan dataset is reconstructed in three dimensions to obtain a three-dimensional model of the shell including the stop and the positioning hole; Extract the three-dimensional coordinate data of the stop and the positioning hole from the three-dimensional model of the housing.

3. The method according to claim 1, characterized in that, The acquisition of the surface three-dimensional geometric data of the ducted fan housing includes: An optical scanning device is used to project an coded speckle pattern onto the surface to be scanned of the ducted fan housing, and an image of the speckle pattern reflected from the surface to be scanned is acquired. Stereo matching is performed on the acquired speckle pattern image to obtain a disparity map of the surface to be scanned; Based on the parallax map and the calibration parameters of the optical scanning device, the three-dimensional point cloud data of the surface to be scanned is calculated to determine the three-dimensional geometric data of the ducted fan housing surface.

4. The method according to claim 1, characterized in that, The step of rigidly registering and fusing the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system includes: At least three non-collinear marking points are provided on the surface of the ducted fan housing; Collect the first coordinates of each of the marked points in a first coordinate system; wherein, the first coordinate system is the coordinate system established when acquiring the three-dimensional coordinate data of the internal structure; The second coordinates of each of the marked points are acquired in a second coordinate system; wherein, the second coordinate system is the coordinate system established when acquiring the three-dimensional geometric data of the surface; Calculate the rigid transformation matrix from the second coordinate system to the first coordinate system based on the first coordinate system and the second coordinate of each of the marked points; The three-dimensional geometric data of the surface is transformed to the first coordinate system by the rigid transformation matrix, and then fused with the three-dimensional coordinate data of the internal structure to obtain a virtual shell model in a unified coordinate system.

5. The method according to claim 4, characterized in that, The step of calculating the rigid transformation matrix from the second coordinate system to the first coordinate system based on the first coordinate and the second coordinate of each of the marked points includes: Calculate the centroid of the first coordinate and the centroid of the second coordinate for each of the marked points to obtain the first centroid coordinates and the second centroid coordinates; Subtract the first centroid coordinate from each of the first coordinates to obtain each of the first decentroid coordinates, and subtract the second centroid coordinate from each of the second coordinates to obtain each of the second decentroid coordinates; Calculate the rotation matrix based on each of the first decentroid coordinates and each of the second decentroid coordinates; Calculate the translation vector based on the first centroid coordinates, the second centroid coordinates, and the rotation matrix; By combining the rotation matrix and the translation vector, a rigid transformation matrix from the second coordinate system to the first coordinate system is obtained.

6. The method according to claim 1, characterized in that, The step of generating a differentiated processing path for the ducted fan housing based on the wall thickness distribution data includes: Obtain the design wall thickness value of the ducted fan housing, and multiply the design wall thickness value by a preset proportional coefficient to obtain the wall thickness threshold. Regions with wall thickness values ​​not less than the wall thickness threshold are identified as thick-walled candidate regions, and regions with wall thickness values ​​less than the wall thickness threshold are identified as thin-walled candidate regions. The strip-shaped areas on both sides of the boundary line between the thick-walled candidate area and the thin-walled candidate area, and whose distance from the boundary line does not exceed a preset transition width, are divided into the third processing area; the area in the thick-walled candidate area other than the third processing area is determined as the first processing area, and the area in the thin-walled candidate area other than the third processing area is determined as the second processing area; Differentiated cutting parameters are assigned to the first processing area, the second processing area, and the third processing area, and differentiated processing paths associated with the cutting parameters of each processing area are generated.

7. The method according to claim 6, characterized in that, Assigning differentiated cutting parameters to the first machining area and the second machining area, including: A first combination of cutting parameters is determined for the first machining area; wherein the first combination of cutting parameters includes at least a first depth of cut and a first feed rate; Multiply the first back depth of cut by the first preset coefficient to obtain the second back depth of cut of the second processing area; The first feed amount is multiplied by the second preset coefficient to obtain the second feed amount of the second processing area; wherein, both the first preset coefficient and the second preset coefficient are less than 1; By combining the second depth of cut and the second feed rate, a second combination of cutting parameters is obtained.

8. The method according to claim 1, characterized in that, The process of performing CNC finishing according to the differentiated machining path also includes: A mapping relationship between thermal diffusion characteristic time and wall thickness is established in advance through calibration; Collect transient temperature data of the machined surface after the tool passes over it; Extract the thermal diffusion characteristic time of the transient temperature data, and invert the remaining wall thickness at the processed surface based on the thermal diffusion characteristic time and the mapping relationship; The cutting parameters for the corresponding machining area are corrected based on the inverted remaining wall thickness.

9. The method according to claim 1, characterized in that, The step of combining the positioning information collected by the infrared positioning device and performing CNC precision machining according to the differentiated machining path includes: At least three non-collinear infrared markers are provided on the ducted fan housing or the clamp that holds the ducted fan housing; The infrared positioning device is used to collect the coordinates of each infrared marker in the machine tool coordinate system in real time, and the actual position and orientation of the ducted fan housing in the machine tool coordinate system are calculated. Calculate the pose deviation between the actual pose and the theoretical pose, and perform rigid body transformation compensation on the differentiated processing path according to the pose deviation; Perform CNC finishing according to the compensated machining path.

10. An infrared-assisted lightweight precision machining system for a ducted fan housing, characterized in that, The system includes: An internal structure acquisition module is used to acquire three-dimensional coordinate data of the internal structure of the ducted fan housing; wherein, the internal structure includes a stop and positioning holes for positioning and installation. The surface scanning module is used to acquire the three-dimensional geometric data of the surface of the ducted fan housing; The registration and fusion module is used to rigidly register and fuse the surface three-dimensional geometric data with the internal structure three-dimensional coordinate data to obtain a virtual shell model in a unified coordinate system. A wall thickness extraction module is used to extract wall thickness distribution data from the virtual shell model; The path generation module is used to generate a differentiated machining path for the ducted fan housing based on the wall thickness distribution data; wherein, the cutting parameters of each machining area on the differentiated machining path are set differently according to the wall thickness distribution of the corresponding machining area. The machining execution module is used to perform CNC precision machining according to the differentiated machining path by combining the positioning information collected by the infrared positioning device.

Images