An aviation part machining process parameter intelligent optimization method and system

Abstract

The application provides an aviation part machining process parameter intelligent optimization method and system, and relates to the technical field of process parameter optimization.The method comprises the following steps: acquiring global geometric appearance point cloud data of a labyrinth seal tooth assembly of an aero-engine rotor after a stabilization heat treatment process; determining a plurality of stress balance compensation inhibition points in the remaining thin-wall tooth shape area, reverse pulse pressure of each stress balance compensation inhibition point in the remaining thin-wall tooth shape area, and eliminating residual stress of the labyrinth seal tooth assembly of the aero-engine rotor based on the reverse pulse pressure of each stress balance compensation inhibition point in the typical thin-wall tooth shape sensitive area and the reverse pulse pressure of each stress balance compensation inhibition point in the remaining thin-wall tooth shape area.The method can efficiently and accurately determine the residual stress distribution of the labyrinth seal tooth assembly of the aero-engine rotor and formulate a reverse pulse pressure process scheme.

Description

Technical Field

[0001] This invention relates to the field of process parameter optimization technology, specifically to an intelligent optimization method and system for the processing parameters of aerospace parts. Background Technology

[0002] The labyrinth seal tooth assembly of an aero-engine rotor is a key component ensuring the engine's sealing performance and operational efficiency, and its machining accuracy directly affects the engine's overall operating performance. In actual production, the aero-engine rotor labyrinth seal tooth assembly requires multiple processes, including forging, rough machining, and stabilization heat treatment. However, due to the typical thin-walled structure of the aero-engine rotor labyrinth seal tooth assembly, its structural rigidity is extremely weak, and unevenly distributed residual stress is easily generated internally after heat treatment. Traditional processes mainly rely on standardized heat treatment specifications for stress control, supplemented by manual, experience-based cutting compensation. However, this approach has significant limitations. Conventional inspection methods can only sample a few discrete points, resulting in limited and sparse inspection coverage, making it difficult to comprehensively capture the drastic gradient changes in the stress field within complex thin-walled geometric regions. This limitation leads to insufficient understanding of the stress distribution. During subsequent finishing processes, the accumulated residual stress is instantly released as material is removed, inducing tooth tip tilting, overall roundness deviations, and torsional deformation, resulting in a high scrap rate for parts. Furthermore, existing stress relief methods lack the ability to precisely intervene in both local sensitive areas and the overall stress field, making it impossible to achieve intelligent and proactive control of the stress state of complex thin-walled components and failing to meet the stringent requirements of modern aerospace manufacturing for high precision and high stability.

[0003] Therefore, how to efficiently and accurately determine the residual stress distribution of the labyrinth seal tooth assembly of an aero-engine rotor and formulate a reverse pulse pressure process scheme is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem solved by this invention is how to efficiently and accurately determine the residual stress distribution of the labyrinth seal tooth assembly of an aero-engine rotor and formulate a reverse pulse pressure process scheme.

[0005] According to a first aspect, the present invention provides an intelligent optimization method for processing parameters of aerospace components, comprising: acquiring point cloud data of the global geometric topography of an aero-engine rotor labyrinth seal tooth assembly after a stabilization heat treatment process; determining multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process; performing X-ray diffraction detection on the multiple basic residual stress detection points of each typical thin-walled tooth profile sensitive region to obtain residual stress detection data of the multiple basic residual stress detection points; performing clustering based on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region; and determining the sensitive region of each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process and the multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region. Multiple supplementary residual stress detection points are used; X-ray diffraction detection is performed on the multiple supplementary residual stress detection points in each typical thin-walled tooth profile sensitive area to obtain residual stress detection data for the multiple supplementary residual stress detection points; based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area are determined; based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area, multiple stress balance compensation suppression points in other thin-walled tooth profile areas and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile areas are determined; based on the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile areas, residual stress is eliminated in the aero-engine rotor labyrinth seal tooth assembly.

[0006] In one possible implementation, determining multiple stress balance compensation suppression points in each typical thin-walled tooth-shaped sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point within the typical thin-walled tooth-shaped sensitive region, based on the residual stress detection data from the multiple basic residual stress detection points and the multiple supplementary residual stress detection points, includes: generating multiple simulated residual stress detection point data based on the residual stress detection data from the multiple basic residual stress detection points and the multiple supplementary residual stress detection points; and constructing a residual stress map, wherein the residual stress map includes multiple nodes and multiple edges between the nodes. The multiple nodes include multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes. The node characteristics of the basic residual stress detection nodes are the residual stress detection data of the basic residual stress detection points, the node characteristics of the supplementary residual stress detection nodes are the residual stress detection data of the supplementary residual stress detection points, and the node characteristics of the simulated residual stress detection nodes are the simulated residual stress detection point data. The residual stress spectrum is processed based on a graph neural network to obtain multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area, and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area.

[0007] In one possible implementation, determining multiple stress balance compensation suppression points in other thin-walled tooth regions and the reverse pulse pressure of each stress balance compensation suppression point in the other thin-walled tooth regions based on multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region includes: determining the residual stress distribution information of other thin-walled tooth regions based on the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly; and determining multiple stress balance compensation suppression points in other thin-walled tooth regions and the reverse pulse pressure of each stress balance compensation suppression point in the other thin-walled tooth regions based on the residual stress distribution information of other thin-walled tooth regions, multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point in the other thin-walled tooth regions.

[0008] In one possible implementation, the step of clustering the residual stress detection data from the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters includes: using the K-means clustering algorithm to cluster the residual stress detection data from the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters.

[0009] According to a second aspect, the present invention provides an intelligent optimization system for processing parameters of aerospace components, comprising: a point cloud data acquisition module, used to acquire point cloud data of the overall geometric shape of an aero-engine rotor labyrinth sealing tooth assembly after a stabilization heat treatment process; a sensitive region determination module, used to determine multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the overall geometric shape of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process; a basic stress detection module, used to perform X-ray diffraction detection on the multiple basic residual stress detection points of each typical thin-walled tooth profile sensitive region to obtain residual stress detection data of multiple basic residual stress detection points; a stress cluster clustering module, used to perform clustering based on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region; and a supplementary detection point determination module, used to determine the sensitive region of each typical thin-walled tooth profile sensitive region based on the point cloud data of the overall geometric shape of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and the multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region. The system comprises: multiple supplementary residual stress detection points in the sensing region; a supplementary stress detection module for performing X-ray diffraction detection on multiple supplementary residual stress detection points in each typical thin-walled tooth profile sensing region to obtain residual stress detection data for multiple supplementary residual stress detection points; a typical suppression parameter determination module for determining multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensing region and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensing region based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points; a remaining suppression parameter determination module for determining multiple stress balance compensation suppression points in other thin-walled tooth profile regions and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile regions based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensing region; and a stress relief execution module for eliminating residual stress in the aero-engine rotor labyrinth seal tooth assembly based on the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensing region and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile regions.

[0010] In one possible implementation, the typical suppression parameter determination module is further configured to: generate multiple simulated residual stress detection point data based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points; construct a residual stress map, the residual stress map including multiple nodes and multiple edges between the multiple nodes, the multiple nodes including multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes, the node characteristics of the basic residual stress detection nodes being the residual stress detection data of the basic residual stress detection points, the node characteristics of the supplementary residual stress detection nodes being the residual stress detection data of the supplementary residual stress detection points, and the node characteristics of the simulated residual stress detection nodes being the simulated residual stress detection point data; process the residual stress map based on a graph neural network to obtain multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive region.

[0011] In one possible implementation, the remaining suppression parameter determination module is further configured to: determine the residual stress distribution information of the remaining thin-walled tooth region based on the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly; and determine the multiple stress balance compensation suppression points and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region based on the residual stress distribution information of the remaining thin-walled tooth region, multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive region.

[0012] In one possible implementation, the stress clustering module is further configured to: use the K-means clustering algorithm to cluster the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters.

[0013] According to a third aspect, embodiments of the present invention provide an electronic device, including: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method described above, the method comprising: acquiring point cloud data of the global geometric topography of an aero-engine rotor labyrinth sealing tooth assembly after a stabilization heat treatment process; determining multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process; determining multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on X-ray diffraction detection of the multiple basic residual stress detection points to obtain residual stress detection data for the multiple basic residual stress detection points; performing clustering based on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region; and performing X-ray diffraction detection on the multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and ... Multiple residual stress distribution clusters in the sensing region are identified, and multiple supplementary residual stress detection points are determined for each typical thin-walled tooth profile sensitive region. X-ray diffraction detection is performed on these supplementary residual stress detection points in each typical thin-walled tooth profile sensitive region to obtain residual stress detection data for each supplementary residual stress detection point. Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple stress balance compensation suppression points and reverse pulse pressures for each stress balance compensation suppression point in each typical thin-walled tooth profile sensitive region are determined. Based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive region, multiple stress balance compensation suppression points and reverse pulse pressures for each stress balance compensation suppression point in other thin-walled tooth profile regions are determined. Based on the reverse pulse pressures for each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive region and the reverse pulse pressures for each stress balance compensation suppression point in other thin-walled tooth profile regions, residual stress is eliminated in the aero-engine rotor labyrinth seal tooth assembly.

[0014] According to the fourth aspect, this embodiment provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the aforementioned intelligent optimization method for aerospace component processing parameters. The method includes: acquiring point cloud data of the global geometric topography of an aero-engine rotor labyrinth sealing tooth assembly after a stabilization heat treatment process; determining multiple typical thin-walled tooth profile sensitive areas and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive area based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process; performing X-ray diffraction detection on the multiple basic residual stress detection points of each typical thin-walled tooth profile sensitive area to obtain residual stress detection data for multiple basic residual stress detection points; performing clustering based on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive area; and performing X-ray diffraction detection on the multiple basic residual stress detection points of each typical thin-walled tooth profile sensitive area based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and the multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive area. The residual stress distribution clusters are used to determine multiple supplementary residual stress detection points in each typical thin-walled tooth profile sensitive area. X-ray diffraction is performed on these supplementary residual stress detection points in each typical thin-walled tooth profile sensitive area to obtain residual stress detection data. Based on the residual stress detection data from the multiple basic residual stress detection points and the multiple supplementary residual stress detection points, multiple stress balance compensation suppression points and the reverse pulse pressure of each stress balance compensation suppression point within each typical thin-walled tooth profile sensitive area are determined. Based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area, multiple stress balance compensation suppression points and the reverse pulse pressure of each stress balance compensation suppression point within the remaining thin-walled tooth profile areas are determined. Based on the reverse pulse pressure of each stress balance compensation suppression point within the typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point within the remaining thin-walled tooth profile areas, residual stress is eliminated in the aero-engine rotor labyrinth seal tooth assembly.

[0015] This invention provides an intelligent optimization method and system for the processing parameters of aerospace components. The method includes: acquiring point cloud data of the global geometric topography of an aero-engine rotor labyrinth sealing tooth assembly after a stabilization heat treatment process; determining multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process; performing X-ray diffraction detection on the multiple basic residual stress detection points of each typical thin-walled tooth profile sensitive region to obtain residual stress detection data for multiple basic residual stress detection points; performing clustering based on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region; determining multiple supplementary residual stress detection points for each typical thin-walled tooth profile sensitive region based on the point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and the multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region; and performing X-ray diffraction detection on the multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region. Multiple supplementary residual stress detection points in the sensitive area are subjected to X-ray diffraction detection to obtain residual stress detection data for multiple supplementary residual stress detection points. Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area are determined. Based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area, multiple stress balance compensation suppression points in other thin-walled tooth profile areas and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile areas are determined. Based on the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in other thin-walled tooth profile areas, the residual stress of the aero-engine rotor labyrinth seal tooth assembly is eliminated. This method can efficiently and accurately determine the residual stress distribution of the aero-engine rotor labyrinth seal tooth assembly and formulate a reverse pulse pressure process scheme. Attached Figure Description

[0016] Figure 1 A flowchart illustrating an intelligent optimization method for machining process parameters of aerospace components provided in an embodiment of the present invention;

[0017] Figure 2 A schematic diagram of an aero-engine rotor labyrinth sealing tooth assembly provided in an embodiment of the present invention;

[0018] Figure 3A flowchart illustrating the process of determining multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive region and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive region, provided for an embodiment of the present invention;

[0019] Figure 4 This is a flowchart illustrating a process for determining multiple stress balance compensation suppression points within the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point within the remaining thin-walled tooth region, as provided in an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of an intelligent optimization system for the processing parameters of aerospace parts, provided as an embodiment of the present invention. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0022] In this embodiment of the invention, the following are provided: Figure 1 The method for intelligent optimization of process parameters in aerospace component manufacturing processes, as shown, includes steps S1 to S9:

[0023] Step S1: Obtain the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process.

[0024] The labyrinth seal tooth assembly of an aero-engine rotor is a critical sealing component. Through the staggered tooth gaps between the aero-engine rotor labyrinth seal tooth assembly and the stator, it creates a throttling effect, thereby achieving multi-stage suppression of airflow leakage. Figure 2 This is a schematic diagram of a labyrinth sealing tooth assembly for an aero-engine rotor, provided as an embodiment of the present invention.

[0025] The global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process refers to the set of three-dimensional coordinate data generated by scanning the labyrinth seal tooth assembly after forging, rough machining and stabilization heat treatment using a high-precision three-dimensional laser scanning device.

[0026] The global geometric point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process contains the three-dimensional coordinate information and spatial distribution characteristics of the assembly surface. This data accurately reflects the minute deformations and geometric contours of the sealing tooth assembly after heat treatment, such as the tooth tip tilt angle, overall roundness deviation, and geometric irregularities in the thickness direction.

[0027] Step S2: Based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process, determine multiple typical thin-walled tooth sensitive areas and multiple basic residual stress detection points for each typical thin-walled tooth sensitive area.

[0028] In some embodiments, a sensitive region determination model can be used to determine multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region. The sensitive region determination model is a convolutional neural network model. The input to the sensitive region determination model is the global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process, and the output of the sensitive region determination model is multiple typical thin-walled tooth profile sensitive regions and multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive region.

[0029] Convolutional Neural Network (CNN) models are a type of multi-layered neural network capable of processing data with a grid-like structure. A CNN consists of convolutional layers, pooling layers, and fully connected layers. CNNs exhibit excellent performance in tasks such as image recognition and geometric feature extraction.

[0030] Several typical thin-walled tooth profile sensitive areas refer to the key geometric areas in labyrinth seal tooth assemblies that are prone to severe deformation or instability due to residual stress release.

[0031] After forging and heat treatment, the labyrinth seal tooth assembly of an aero-engine rotor exhibits weak structural rigidity and significant residual stress concentration. This makes it highly susceptible to warping, twisting, and out-of-tolerance tooth tip precision during subsequent finishing processes due to uneven release of internal residual stress. Typical sensitive areas for thin-walled tooth profiles can be located at the thin-walled tooth tip, the transition area of ​​the tooth root arc, and the radial section most significantly affected by heat treatment. These sensitive areas typically have small tooth wall thickness, high stress concentration, and relatively weak structural rigidity.

[0032] Multiple basic residual stress detection points in each typical thin-walled tooth profile sensitive area are points distributed within each typical thin-walled tooth profile sensitive area for preliminary detection of residual stress.

[0033] The global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process provides complete spatial geometric coordinates and dense surface sampling points, which include micro-morphological feature changes caused by internal residual stress. These subtle deviations in morphology can reflect the trend of internal stress concentration and potential instability.

[0034] Convolutional neural networks (CNNs) utilize their deep convolutional kernels to extract multi-scale features from the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after stabilization heat treatment. The convolutional layers first capture local geometric gradient features in the point cloud data, such as the magnitude of the change in the tooth wall normal vector and the curvature distribution at the tooth tip. Through multi-layer nonlinear mapping, the CNN can transform these geometric features into physical stability evaluation indicators, then identify the regions with the weakest rigidity and largest geometric deviation in the assembly as several typical thin-walled tooth profile sensitive areas. Next, based on the predicted stress field gradient values ​​within the sensitive areas, the model performs spatial sampling at key nodes of stress concentration in the geometric profile, thereby accurately calculating and calibrating multiple basic residual stress detection points for each typical thin-walled tooth profile sensitive area.

[0035] Step S3: Perform X-ray diffraction detection on multiple basic residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data for multiple basic residual stress detection points.

[0036] Residual stress detection data from multiple foundation residual stress detection points refers to the residual stress measurement data obtained after irradiating multiple foundation residual stress detection points with an X-ray diffractometer and analyzing the diffraction peak displacements. The residual stress detection data for each foundation residual stress detection point includes the principal stress direction, shear stress value, and surface residual compressive stress / tensile stress value at each detection point location.

[0037] Step S4: Cluster the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth-shaped sensitive area.

[0038] In some embodiments, the residual stress detection data of the multiple basic residual stress detection points can be clustered using the K-means clustering algorithm to obtain multiple residual stress distribution clusters.

[0039] K-means clustering is an unsupervised clustering algorithm. Through iterative calculation, K-means clustering divides data into K clusters, minimizing the differences between data points within each cluster and maximizing the differences between data points in different clusters.

[0040] Multiple residual stress distribution clusters in a typical thin-walled tooth-shaped sensitive area are obtained by using the K-means clustering algorithm to cluster the residual stress detection data of each detection point in the sensitive area according to the correlation between stress characteristics and spatial location, resulting in a data set with similar stress distribution characteristics. Each cluster represents a stress level state; for example, multiple residual stress distribution clusters may include high-stress tensile clusters, medium-stress equilibrium clusters, and low-stress compressive clusters.

[0041] Each typical thin-walled tooth profile sensitive region contains multiple clusters of residual stress distributions. The number of residual stress distribution clusters within each typical thin-walled tooth profile sensitive region corresponds to the K value of the K-means clustering algorithm.

[0042] In some embodiments, the value of K can be determined by a preset relationship table between the value of K and the spatial distribution density of detection points within a typical thin-walled tooth profile sensitive area. The higher the spatial distribution density of detection points within a typical thin-walled tooth profile sensitive area, the larger the value of K. The preset relationship table between the value of K and the spatial distribution density of detection points within a typical thin-walled tooth profile sensitive area is artificially constructed in advance.

[0043] In some embodiments, the process of clustering residual stress detection data from multiple basic residual stress detection points using the K-means clustering algorithm is as follows: The algorithm first maps the residual stress detection data from multiple basic residual stress detection points to a numerical feature space. Then, it randomly initializes cluster centers in the numerical feature space and calculates the feature distance between the stress feature of each basic residual stress detection point and the center point. Next, based on the proximity of the stress values ​​and the spatial distance of the detection points on the sealing tooth, stress data points with similar attributes are grouped into the same set. By continuously updating the cluster centers until the change in the cluster centers is less than a preset threshold, each resulting cluster is a residual stress distribution cluster, and each residual stress distribution cluster represents a typical residual stress state within a sensitive area of ​​a typical thin-walled tooth profile.

[0044] By clustering the residual stress detection data from multiple basic residual stress detection points, the originally discrete detection points can be transformed into physically meaningful stress distribution patches, thereby clarifying the coverage and distribution boundaries of different stress energy levels within the sensitive area.

[0045] Step S5: Based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and the multiple residual stress distribution clusters of each typical thin-walled tooth sensitive area, determine multiple supplementary residual stress detection points for each typical thin-walled tooth sensitive area.

[0046] In some embodiments, an augmented detection analysis model can be used to determine multiple augmented residual stress detection points for each typical thin-walled tooth profile sensitive region. The augmented detection analysis model is a convolutional neural network. The input to the augmented detection analysis model is the global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly after the stabilization heat treatment process, and multiple residual stress distribution clusters for each typical thin-walled tooth profile sensitive region. The output of the augmented detection analysis model is multiple augmented residual stress detection points for each typical thin-walled tooth profile sensitive region.

[0047] Multiple supplementary residual stress detection points in each typical thin-walled tooth-shaped sensitive area were determined by the supplementary detection analysis model. These additional detection locations were determined based on the preliminary detection to improve the accuracy of stress field simulation.

[0048] Multiple supplementary residual stress detection points in the sensitive region of a typical thin-walled tooth profile are distributed in the transition zone where the stress distribution gradient changes drastically as obtained from cluster analysis, and can be used to perform more detailed blinding and correction of complex stress fields.

[0049] The point cloud data of the global geometric topography of the labyrinth seal tooth assembly of an aero-engine rotor after the stabilization heat treatment process can represent the geometric feature information of each sampling location within the sensitive area of ​​a typical thin-walled tooth profile. Multiple residual stress distribution clusters within each typical thin-walled tooth profile sensitive area can reflect the distribution pattern of residual stress at different locations within the area. By combining geometric features and residual stress distribution patterns, the model can identify the transition regions between clusters where residual stress changes drastically. These regions require more detection points to accurately describe the residual stress distribution.

[0050] Convolutional neural networks (CNNs) can fuse stress distribution information corresponding to multiple residual stress distribution clusters within a typical thin-walled tooth profile sensitive area. CNNs can extract the geometric features of the sensitive area and analyze the correlation between different residual stress distribution clusters and local geometric features. Then, by calculating the gradient of residual stress change within the sensitive area, CNNs can identify regions with drastic residual stress changes and where no basic residual stress detection points have been set. Finally, CNNs can select key locations in these regions as supplementary residual stress detection points to improve the accuracy of residual stress distribution description.

[0051] In some embodiments, determining multiple supplementary residual stress detection points for each typical thin-walled tooth sensitive region based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and multiple residual stress distribution clusters for each typical thin-walled tooth sensitive region includes steps S21 to S23:

[0052] Step S21: Based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and multiple residual stress distribution clusters of each typical thin-walled tooth sensitive area, determine the surface topography and stress field correlation distribution map of each typical thin-walled tooth sensitive area, the stress field change gradient between adjacent residual stress distribution clusters, and the stress state similarity between adjacent residual stress distribution clusters.

[0053] In some embodiments, a convolutional neural network can be used to determine the surface morphology and stress field correlation distribution map of each typical thin-walled tooth-shaped sensitive region, the stress field change gradient between adjacent residual stress distribution clusters, and the stress state similarity between adjacent residual stress distribution clusters.

[0054] The distribution map of the correlation between the surface morphology and stress field of a typical thin-walled tooth-shaped sensitive area is a spatial distribution image that reflects the correspondence between the surface geometry and the residual stress accumulation trend of a typical thin-walled tooth-shaped sensitive area.

[0055] The correlation diagram between the surface morphology and stress field of a typical thin-walled tooth-shaped sensitive area represents the probability of a high residual stress level at the corresponding geometric location on the typical thin-walled tooth-shaped sensitive area.

[0056] The stress field variation gradient between adjacent residual stress distribution clusters refers to a numerical index of the degree of drastic change in stress magnitude between adjacent residual stress distribution clusters.

[0057] The stress state similarity between adjacent residual stress distribution clusters is a numerical index that represents the degree of similarity in stress distribution states between adjacent residual stress distribution clusters.

[0058] Convolutional neural networks (CNNs) can extract spatial features from the global geometric topography point cloud data of aero-engine rotor labyrinth sealing tooth assemblies after stabilization heat treatment, and identify geometric structural information such as protrusions and depressions, and wall thickness variations on the sealing tooth surface. CNNs can perform numerical distribution analysis on multiple residual stress distribution clusters in sensitive areas of each typical thin-walled tooth profile, while simultaneously learning the stress patterns and spatial distribution patterns within different clusters. The model can fuse and infer geometric topography information with stress distribution information to calculate the correlation between surface topography and stress field, thereby obtaining a correlation distribution map of surface topography and stress field for each sensitive area of ​​a typical thin-walled tooth profile. By comparing the stress values ​​and distribution patterns of adjacent residual stress distribution clusters, the model can obtain the stress field change gradient between adjacent residual stress distribution clusters. Furthermore, the model can quantitatively evaluate the overall distribution characteristics of adjacent residual stress distribution clusters to determine the stress state similarity between adjacent residual stress distribution clusters.

[0059] Step S22: Based on the multiple residual stress distribution clusters of each typical thin-walled tooth profile sensitive region, the surface morphology and stress field correlation distribution map of each typical thin-walled tooth profile sensitive region, the stress field change gradient between adjacent residual stress distribution clusters, and the stress state similarity between adjacent residual stress distribution clusters, determine the set of stress blind zone locations and the range of residual stress abrupt change boundary within each typical thin-walled tooth profile sensitive region.

[0060] In some embodiments, a convolutional neural network can be used to determine the set of stress blind zone locations for each typical thin-walled tooth sensitive region and the range of residual stress abrupt change boundaries within each typical thin-walled tooth sensitive region.

[0061] The set of stress blind zone locations in a typical thin-walled tooth profile sensitive area is a set of local locations within the typical thin-walled tooth profile sensitive area that cannot be covered by existing residual stress detection points and for which stress state information is unknown.

[0062] The residual stress abrupt change boundary range within each typical thin-walled tooth profile sensitive area is a continuous coordinate range consisting of the locations where the stress value within each typical thin-walled tooth profile sensitive area changes drastically.

[0063] Convolutional neural networks (CNNs) can identify sets of stress blind spots where the stress state cannot be determined and the points covered by multiple residual stress distribution clusters in the sensitive region of each typical thin-walled tooth profile can be analyzed. This can be achieved by combining the coverage of these clusters with the correlation between surface morphology and stress field distribution maps. Furthermore, by combining the stress field gradient between adjacent residual stress distribution clusters with the stress state similarity between them, CNNs can analyze the boundaries of regions with significant stress value changes, thereby determining the range of abrupt residual stress change boundaries within the sensitive region of each typical thin-walled tooth profile.

[0064] Step S23: Based on the surface morphology and stress field correlation distribution map of each typical thin-walled tooth sensitive area, the set of stress blind zone locations of each typical thin-walled tooth sensitive area, and the range of residual stress abrupt change boundary within each typical thin-walled tooth sensitive area, determine multiple supplementary residual stress detection points for each typical thin-walled tooth sensitive area.

[0065] In some embodiments, a convolutional neural network can be used to determine multiple supplementary residual stress detection points for each typical thin-walled tooth-shaped sensitive region.

[0066] Convolutional neural networks (CNNs) can identify key locations with high correlation strength by numerically analyzing the correlation distribution between the surface morphology and stress field of each typical thin-walled tooth-shaped sensitive region. Then, the CNN spatially fuses the set of stress blind spots with the boundary range of residual stress abrupt changes, thereby filtering out locations that possess both data gaps and high stress variability. Under the premise of meeting the requirements of data acquisition coverage and stress characterization accuracy, the CNN can selectively filter candidate locations, ultimately determining multiple supplementary residual stress detection points for each typical thin-walled tooth-shaped sensitive region.

[0067] Step S6: Perform X-ray diffraction detection on multiple supplementary residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data for multiple supplementary residual stress detection points.

[0068] The residual stress detection data from multiple supplementary residual stress detection points refers to the residual stress measurement data obtained after irradiating multiple supplementary residual stress detection points with an X-ray diffractometer and analyzing the diffraction peak displacements. The residual stress detection data for each supplementary residual stress detection point includes the principal stress direction, shear stress value, and surface residual compressive stress / tensile stress value at each detection point location.

[0069] The residual stress detection data from multiple additional residual stress detection points can supplement the measured residual stress information in areas with drastic changes in residual stress gradient within the sensitive region, and improve the data integrity of the entire residual stress field in the sensitive region of a typical thin-walled tooth profile, thereby enhancing the accuracy of the residual stress distribution description.

[0070] Step S7: Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, determine multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth profile sensitive area.

[0071] In some embodiments, Figure 3This invention provides a schematic flowchart for determining multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive region and the reverse pulse pressure of each stress balance compensation suppression point within the typical thin-walled tooth profile sensitive region. The determination of multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive region and the reverse pulse pressure of each stress balance compensation suppression point within the typical thin-walled tooth profile sensitive region includes steps S31 to S33:

[0072] Step S31: Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, generate multiple simulated residual stress detection point data.

[0073] In some embodiments, a residual stress simulation model can be used to generate data from multiple simulated residual stress detection points. The residual stress simulation model is a variational autoencoder. The inputs to the residual stress simulation model are the residual stress detection data from the multiple basic residual stress detection points and the residual stress detection data from the multiple supplementary residual stress detection points; the output of the residual stress simulation model is the data from multiple simulated residual stress detection points.

[0074] Variational autoencoders (VACs) are generative models consisting of an encoder and a decoder. The encoder compresses high-dimensional input data into a low-dimensional latent space distribution, thereby learning the underlying statistical regularities. The decoder samples from the latent space and reconstructs new samples consistent with the original data distribution. By introducing variational inference, VACs not only learn data reconstruction but also possess powerful generative capabilities. VACs can infer the complete data distribution based on a limited number of observations and have significant advantages in data augmentation and field simulation tasks.

[0075] Multiple simulated residual stress detection point data are non-measured point residual stress simulation measurement data generated by a residual stress simulation model in a typical thin-walled tooth-shaped sensitive area to complete the stress field distribution. Each simulated residual stress detection point data includes the predicted principal stress value, predicted shear stress value, and predicted surface residual compressive / tensile stress value for each simulated point location.

[0076] Residual stress detection data from multiple basic residual stress detection points characterize the true stress state of the basic measuring points within the sensitive area and reflect the overall distribution profile of residual stress inside the part. Residual stress detection data from multiple supplementary residual stress detection points further refine the characterization of local stress characteristics in areas with drastic changes in residual stress gradient. These detection data cover different residual stress states and geometric feature locations within the sensitive area of ​​a typical thin-walled tooth profile, and comprehensively reflect the overall distribution law of residual stress within the area. Using this detection data as model learning samples, the model can learn and identify the nonlinear correlation between different geometric locations and residual stress values.

[0077] The variational autoencoder first transforms residual stress detection data from multiple basic and supplementary residual stress detection points into a statistically significant latent variable distribution. The model then learns the spatial correlation and numerical distribution characteristics of residual stress in the thin-walled structure from this distribution. The model utilizes a self-attention mechanism to analyze the mapping relationship between spatial distances and stress differences between detection points, constructing a continuous stress field probability distribution within the latent space. Subsequently, the decoder performs high-density sampling along the trajectory of stress field gradient changes in the latent space and remaps these sampling points back to the geometric coordinates of the labyrinth sealing tooth assembly. During the generation process, the model can smooth and correct the sampling results, ensuring that the generated points fill the gaps between measured points while maintaining the continuity of the stress field. Finally, through nonlinear transformations of multiple layers of neurons, the model can transform the latent variables into stress predictions at specific locations, thereby generating multiple simulated residual stress detection point data.

[0078] Step S32: Construct a residual stress map. The residual stress map includes multiple nodes and multiple edges between the nodes. The multiple nodes include multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes. The node characteristics of the basic residual stress detection nodes are the residual stress detection data of the basic residual stress detection points. The node characteristics of the supplementary residual stress detection nodes are the residual stress detection data of the supplementary residual stress detection points. The node characteristics of the simulated residual stress detection nodes are the simulated residual stress detection point data.

[0079] The residual stress map can express the geometric adjacency relationships between different detection locations and the physical logic of stress transfer. The residual stress map includes multiple nodes and edges. The nodes include basic residual stress detection nodes, supplementary residual stress detection nodes, and simulated residual stress detection nodes. Each node and edge has specific characteristics.

[0080] The residual stress detection nodes represent the residual stress detection points distributed within the sensitive area of ​​a typical thin-walled tooth profile. The node characteristics are the residual stress detection data of the residual stress detection points.

[0081] The supplementary residual stress detection nodes represent the supplementary residual stress detection points distributed in the stress gradient transition zone of the sensitive area of ​​a typical thin-walled tooth profile. The node characteristics are the residual stress detection data of the supplementary residual stress detection points.

[0082] The simulated residual stress detection node represents the simulated residual stress detection point, and its node characteristics are the simulated residual stress detection point data.

[0083] An edge represents the spatial relationship between any two nodes, and the characteristics of an edge are the positional relationship information between the nodes.

[0084] Location relationship information specifically includes the spatial distance and relative direction between nodes.

[0085] The residual stress distribution in the labyrinth seal tooth assembly of an aero-engine rotor exhibits spatial continuity. By constructing a residual stress map, the spatial positional relationship and stress characteristic correlation between different types of detection nodes can be clearly characterized, thereby achieving a structured and intuitive presentation of the global residual stress field in the sensitive area of ​​a typical thin-walled tooth profile.

[0086] Step S33: Process the residual stress spectrum based on the graph neural network to obtain multiple stress balance compensation and suppression points in each typical thin-walled tooth sensitive region and the reverse pulse pressure of each stress balance compensation and suppression point in the typical thin-walled tooth sensitive region.

[0087] Graph Neural Networks (GNNs) are deep learning models specifically designed for processing graph data. GNNs can directly define convolution operations on graphs in non-Euclidean space. Through message passing mechanisms, GNNs enable each node to continuously aggregate feature information from its neighbors, thereby capturing local and even global topological dependencies. GNNs can abstract high-dimensional physical representations from complex node and edge features, thus enabling prediction of specific physical quantities or key nodes. The input to the GNN is the residual stress spectrum, and the output is multiple stress balance compensation suppression points in each typical thin-walled tooth-shaped sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point within the typical thin-walled tooth-shaped sensitive region.

[0088] Multiple stress balance compensation suppression points in the sensitive region of each typical thin-walled tooth profile were determined through graph neural network analysis to be the optimal physical locations for applying reverse pulse pressure to eliminate residual stress.

[0089] Multiple stress balance compensation and suppression points in the sensitive region of each typical thin-walled tooth profile are distributed at key locations that can have the greatest intervention effect on the internal stress field of the labyrinth seal tooth assembly of an aero-engine.

[0090] The reverse pulse pressure at each stress balance compensation suppression point within a typical thin-walled tooth-shaped sensitive region is the magnitude of the pulse pressure applied to each stress balance compensation suppression point to counteract the residual stress in that region.

[0091] Reverse pulse pressure can be used to counteract residual stress. By generating high-frequency longitudinal waves through pulse pressure and transmitting them into the metal, the originally tight and disordered metal lattice dislocations are induced to slip and rearrange, thereby adjusting the lattice structure and releasing residual stress. By applying this reverse pulse pressure at the corresponding suppression point, the local stress state can be precisely controlled, preventing deformation and precision deviations in parts due to residual stress, and ensuring the stability of the component structure.

[0092] In the residual stress map, each node corresponds to a residual stress detection point, and the node features are the residual stress detection data or simulated residual stress detection data for that point. These features directly reflect the residual stress distribution in the sensitive area of ​​a typical thin-walled tooth profile. The edges between nodes represent the spatial relationship between different points, and the edge features are the positional relationship information between the nodes. This positional relationship information reflects the mutual transmission and influence of residual stress in different regions. By processing the residual stress map using a graph neural network, stress concentration areas can be identified, and stress balance compensation suppression points and their corresponding reverse pulse pressures can be determined. This helps to accurately implement residual stress control and suppress deformation and accuracy deviations of parts after heat treatment.

[0093] Graph neural networks can fuse information from multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes in the residual stress map. In each layer of graph convolution, nodes absorb stress vector information from surrounding nodes according to edge weights, thereby analyzing the imbalance of residual stress distribution globally. The graph neural network identifies the stress concentration areas that contribute most to structural deformation by calculating the local stress potential energy distribution of each node. Then, the model uses a classification prediction branch to filter out geometric nodes in the residual stress map that can effectively suppress deformation, marking them as multiple stress balance compensation suppression points for each typical thin-walled toothed sensitive region. Subsequently, the model uses a regression prediction branch, combined with the material thickness characteristics of the node's location and the resultant stress vector of the adjacent region, to calculate the energy threshold required to offset the residual stress at that location. The graph neural network can simulate the propagation and superposition effects of pulse waves within metal, and continuously optimizes pressure parameters through a backpropagation algorithm, ensuring that the calculated pressure value accurately neutralizes internal stress without causing additional geometric deformation.

[0094] Step S8: Based on the multiple stress balance compensation suppression points in each typical thin-walled tooth profile sensitive area, determine the multiple stress balance compensation suppression points in the remaining thin-walled tooth profile areas and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth profile areas.

[0095] In some embodiments, Figure 4 This is a schematic flowchart illustrating the process of determining multiple stress balance compensation suppression points and the reverse pulse pressure of each stress balance compensation suppression point within the remaining thin-walled tooth region, as provided in an embodiment of the present invention. The determination of these points includes steps S41-S42.

[0096] Step S41: Based on the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly, determine the residual stress distribution information of the remaining thin-walled tooth regions.

[0097] In some embodiments, a residual stress calculation model can be used to determine the residual stress distribution information of the remaining thin-walled tooth region. The residual stress calculation model is a graph convolutional network. The inputs to the residual stress calculation model are the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly; the output of the residual stress calculation model is the residual stress distribution information of the remaining thin-walled tooth region.

[0098] Graph Convolutional Networks (GCNs) are a class of deep learning models designed for spatial data. GCNs leverage spatial topological relationships as their computational foundation, and through neighborhood information aggregation and spatial convolution operations, they can model the inherent dependencies within the data, thereby enabling feature learning and target inference.

[0099] The residual stress distribution information of the remaining thin-walled tooth regions was determined by the residual stress calculation model. This information covers the residual stress state of all thin-walled tooth regions on the aero-engine rotor labyrinth seal tooth assembly, excluding the typical thin-walled tooth sensitive areas. The residual stress distribution information of the remaining thin-walled tooth regions includes the principal stress value, shear stress value, and surface residual compressive / tensile stress value at each point cloud sampling point in the remaining thin-walled tooth regions.

[0100] The residual stress map provides high-precision stress samples and physical transmission laws for typical sensitive areas. The global geometric topography point cloud data of the aero-engine rotor labyrinth seal tooth assembly reflects the microscopic deformation characteristics and geometric contour changes across the entire part. These data reveal the intrinsic relationship between geometric features and residual stress, enabling the model to generalize its representation to unknown areas using the patterns of known sensitive areas.

[0101] Graph convolutional networks (GCNNs) can use the stress-geometric mapping learned from residual stress maps as global constraints. Then, through multi-layer graph convolutions, they aggregate neighborhood features, thereby uncovering the geometric similarities between other thin-walled tooth regions and typical thin-walled tooth sensitive regions. The model learns the patterns of subtle surface geometric distortions caused by stress within the sensitive region and can then transfer these patterns to the morphological features of other thin-walled tooth regions. If the point cloud features of other thin-walled tooth regions exhibit similar geometric trends as the sensitive region, such as tooth tip tilting or roundness deviation, the GCNN can infer the corresponding stress distribution features of that region based on the learned physical mapping weights. Subsequently, through cross-regional node feature fusion, the model can unify discrete geometric features and stress features into a continuous stress field distribution, ultimately outputting the residual stress distribution information of the remaining thin-walled tooth regions.

[0102] Step S42: Based on the residual stress distribution information of the remaining thin-walled tooth region, the multiple stress balance compensation suppression points of each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive region, determine the multiple stress balance compensation suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region.

[0103] In some embodiments, a compensation pressure determination model can be used to determine multiple stress balance compensation suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region. The compensation pressure determination model is a deep neural network. The input to the compensation pressure determination model is the residual stress distribution information of the remaining thin-walled tooth region, multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive region. The output of the compensation pressure determination model is the multiple stress balance compensation suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region.

[0104] Deep Neural Networks (DNNs) are neural network architectures composed of multiple hidden layers. Through complex nonlinear mappings of input features by each neuron in each layer, DNNs can learn deep-seated abstract features and logical relationships within data. DNNs possess powerful generalization and fitting capabilities, and can handle complex mapping problems between multivariate inputs and multi-objective outputs.

[0105] The multiple stress balance compensation suppression points in the remaining thin-walled tooth region were determined by the compensation pressure determination model, and the points in the remaining thin-walled tooth region were used to apply reverse pulse pressure to eliminate residual stress.

[0106] The reverse pulse pressure at each stress balance compensation suppression point in the remaining thin-walled tooth region is determined by the compensation pressure determination model, which is the magnitude of the pulse pressure applied to each stress balance compensation suppression point to counteract residual stress.

[0107] The residual stress distribution information in the remaining thin-walled tooth region can characterize the initial internal stress state of each point cloud sampling point in the non-sensitive area, serving as the basic data for the model's energy balance calculation. Multiple stress balance compensation and suppression points in each typical thin-walled tooth sensitive region, along with the reverse pulse pressure at each stress balance compensation and suppression point within the sensitive region, act as spatial constraints and boundary load conditions during the calculation. Since the aero-engine rotor labyrinth sealing tooth assembly is a continuous thin-walled structure with a high degree of structural stress correlation, the preset pressure load within the sensitive region will generate a stress redistribution effect through the material's interior. Based on the aforementioned stress transmission characteristics and input parameter constraints, the model can deduce the control parameters that can generate a structural stress synergy effect in the remaining thin-walled tooth regions based on the determined energy input state of the sensitive region, ensuring that the residual stress across the entire assembly is globally offset through the superposition of pressure fields at multiple points.

[0108] Deep neural networks can analyze the magnitude, gradient, and distribution trend of stress in various locations within the thin-walled toothed region based on the residual stress distribution information of other thin-walled toothed regions. By combining the spatial distribution of stress balance compensation and suppression points in the sensitive areas of typical thin-walled toothed regions with their reverse pulse pressure parameters, the model can analyze the stress transmission path and load influence range of the entire sealing tooth assembly. Deep neural networks establish a global stress balance relationship through multi-layer nonlinear operations, enabling the calculation of stress interference and coupling patterns between different locations. Based on the goal of minimizing the overall structural energy, deep neural networks can locate the optimal action points in the remaining thin-walled toothed regions to achieve overall stress balance, and calculate the required reverse pulse pressure magnitude at each point based on the existing pulse pressure parameters in the sensitive areas. This allows the pressure in each region to achieve a synergistic cancellation effect across the entire domain, ultimately outputting stress balance compensation and suppression points adapted to the remaining thin-walled toothed regions and their corresponding reverse pulse pressures.

[0109] Step S9: Based on the reverse pulse pressure of each stress balance compensation and suppression point in the sensitive area of ​​the typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile areas, the residual stress of the aero-engine rotor labyrinth seal tooth assembly is eliminated.

[0110] Once the reverse pulse pressure of each stress balance compensation and suppression point in the sensitive area of ​​a typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile areas are determined, the pulse pressure generating device is controlled to apply a pulse pressure of a specified intensity at the corresponding stress balance compensation and suppression point based on the reverse pulse pressure of each stress balance compensation and suppression point in the sensitive area of ​​the typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile areas. Through the dislocation movement and microstructure rearrangement generated by the mechanical wave inside the metal, the precise neutralization and elimination of residual stress in the entire domain of the aero-engine rotor labyrinth sealing tooth assembly is finally achieved.

[0111] Based on the same inventive concept Figure 5 This is a schematic diagram of an intelligent optimization system for the processing parameters of aerospace parts, provided in an embodiment of the present invention. The intelligent optimization system for the processing parameters of aerospace parts includes:

[0112] The point cloud data acquisition module 51 is used to acquire the global geometric shape point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process.

[0113] Sensitive area determination module 52 is used to determine multiple typical thin-walled tooth sensitive areas and multiple basic residual stress detection points for each typical thin-walled tooth sensitive area based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process.

[0114] The basic stress detection module 53 is used to perform X-ray diffraction detection on multiple basic residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data of multiple basic residual stress detection points.

[0115] The stress clustering module 54 is used to cluster the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth-shaped sensitive area.

[0116] The supplementary detection point determination module 55 is used to determine multiple supplementary residual stress detection points for each typical thin-walled tooth sensitive area based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and multiple residual stress distribution clusters of each typical thin-walled tooth sensitive area.

[0117] The supplementary stress detection module 56 is used to perform X-ray diffraction detection on multiple supplementary residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data of multiple supplementary residual stress detection points.

[0118] The typical suppression parameter determination module 57 is used to determine multiple stress balance compensation suppression points and reverse pulse pressure of each stress balance compensation suppression point in each typical thin-walled tooth sensitive area based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points.

[0119] The remaining suppression parameter determination module 58 is used to determine multiple stress balance compensation suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region based on multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive region.

[0120] The stress relief execution module 59 is used to eliminate residual stress in the aero-engine rotor labyrinth seal tooth assembly based on the reverse pulse pressure of each stress balance compensation suppression point in the sensitive area of ​​the typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth profile areas.

[0121] It should be noted that, in order to simplify the descriptions disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments of this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0122] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A method for intelligent optimization of machining process parameters for aerospace components, characterized in that, include: Acquire point cloud data of the global geometric topography of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process; Based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the aforementioned stabilization heat treatment process, multiple typical thin-walled tooth sensitive areas and multiple basic residual stress detection points for each typical thin-walled tooth sensitive area are determined. X-ray diffraction was performed on multiple basic residual stress detection points in each typical thin-walled tooth-shaped sensitive region to obtain residual stress detection data for multiple basic residual stress detection points. Clustering is performed on the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth-shaped sensitive region. Based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process, and multiple residual stress distribution clusters in each typical thin-walled tooth sensitive area, multiple supplementary residual stress detection points are determined for each typical thin-walled tooth sensitive area. X-ray diffraction was performed on multiple supplementary residual stress detection points in each typical thin-walled tooth-shaped sensitive region to obtain residual stress detection data for multiple supplementary residual stress detection points. Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive area are determined. Based on the multiple stress balance compensation and suppression points in each typical thin-walled tooth profile sensitive region, the multiple stress balance compensation and suppression points in the remaining thin-walled tooth profile regions and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile regions are determined. Based on the reverse pulse pressure of each stress balance compensation and suppression point in the sensitive area of ​​the typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile areas, residual stress is eliminated in the labyrinth seal tooth assembly of the aero-engine rotor.

2. The intelligent optimization method for processing parameters of aerospace parts as described in claim 1, characterized in that, The determination of multiple stress balance compensation and suppression points in each typical thin-walled tooth sensitive region based on the residual stress detection data from the multiple basic residual stress detection points and the residual stress detection data from the multiple supplementary residual stress detection points includes: Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple simulated residual stress detection point data are generated. A residual stress map is constructed, which includes multiple nodes and multiple edges between the nodes. The multiple nodes include multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes. The node characteristics of the basic residual stress detection nodes are the residual stress detection data of the basic residual stress detection points. The node characteristics of the supplementary residual stress detection nodes are the residual stress detection data of the supplementary residual stress detection points. The node characteristics of the simulated residual stress detection nodes are the simulated residual stress detection point data. The residual stress spectrum is processed based on a graph neural network to obtain multiple stress balance compensation and suppression points in each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation and suppression point in the typical thin-walled tooth sensitive region.

3. The intelligent optimization method for processing parameters of aerospace parts as described in claim 1, characterized in that, The determination of multiple stress balance compensation and suppression points in other thin-walled tooth regions based on multiple stress balance compensation and suppression points in each typical thin-walled tooth region, and the reverse pulse pressure of each stress balance compensation and suppression point in the other thin-walled tooth regions, includes: Based on the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly, the residual stress distribution information of the remaining thin-walled tooth region is determined. Based on the residual stress distribution information of the remaining thin-walled tooth region, the multiple stress balance compensation and suppression points of each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation and suppression point in the typical thin-walled tooth sensitive region, the multiple stress balance compensation and suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth region are determined.

4. The intelligent optimization method for processing parameters of aerospace parts as described in claim 1, characterized in that, The residual stress detection data based on the multiple basic residual stress detection points are clustered to obtain multiple residual stress distribution clusters, including: The residual stress detection data from the multiple basic residual stress detection points were clustered using the K-means clustering algorithm to obtain multiple residual stress distribution clusters.

5. An intelligent optimization system for machining process parameters of aerospace parts, characterized in that, include: The point cloud data acquisition module is used to acquire the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process. The sensitive area determination module is used to determine multiple typical thin-walled tooth sensitive areas and multiple basic residual stress detection points for each typical thin-walled tooth sensitive area based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process. The basic stress detection module is used to perform X-ray diffraction detection on multiple basic residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data for multiple basic residual stress detection points. The stress clustering module is used to cluster the residual stress detection data of the multiple basic residual stress detection points to obtain multiple residual stress distribution clusters for each typical thin-walled tooth-shaped sensitive region. The supplementary detection point determination module is used to determine multiple supplementary residual stress detection points for each typical thin-walled tooth sensitive area based on the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly after the stabilization heat treatment process and multiple residual stress distribution clusters of each typical thin-walled tooth sensitive area. The supplementary stress detection module is used to perform X-ray diffraction detection on multiple supplementary residual stress detection points in each typical thin-walled tooth-shaped sensitive area to obtain residual stress detection data for multiple supplementary residual stress detection points. The typical suppression parameter determination module is used to determine multiple stress balance compensation suppression points in each typical thin-walled tooth sensitive area and the reverse pulse pressure of each stress balance compensation suppression point in the typical thin-walled tooth sensitive area based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points. The remaining suppression parameter determination module is used to determine multiple stress balance compensation suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation suppression point in the remaining thin-walled tooth region based on multiple stress balance compensation suppression points in the sensitive region of each typical thin-walled tooth. The stress relief execution module is used to eliminate residual stress in the aero-engine rotor labyrinth seal tooth assembly based on the reverse pulse pressure of each stress balance compensation and suppression point in the sensitive area of ​​the typical thin-walled tooth profile and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth profile areas.

6. The intelligent optimization system for aerospace component processing parameters as described in claim 5, characterized in that, The typical suppression parameter determination module is also used for: Based on the residual stress detection data of the multiple basic residual stress detection points and the residual stress detection data of the multiple supplementary residual stress detection points, multiple simulated residual stress detection point data are generated. A residual stress map is constructed, which includes multiple nodes and multiple edges between the nodes. The multiple nodes include multiple basic residual stress detection nodes, multiple supplementary residual stress detection nodes, and multiple simulated residual stress detection nodes. The node characteristics of the basic residual stress detection nodes are the residual stress detection data of the basic residual stress detection points. The node characteristics of the supplementary residual stress detection nodes are the residual stress detection data of the supplementary residual stress detection points. The node characteristics of the simulated residual stress detection nodes are the simulated residual stress detection point data. The residual stress spectrum is processed based on a graph neural network to obtain multiple stress balance compensation and suppression points in each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation and suppression point in the typical thin-walled tooth sensitive region.

7. The intelligent optimization system for aerospace component processing parameters as described in claim 5, characterized in that, The remaining suppression parameter determination module is also used for: Based on the residual stress map and the global geometric topography point cloud data of the aero-engine rotor labyrinth sealing tooth assembly, the residual stress distribution information of the remaining thin-walled tooth region is determined. Based on the residual stress distribution information of the remaining thin-walled tooth region, the multiple stress balance compensation and suppression points of each typical thin-walled tooth sensitive region, and the reverse pulse pressure of each stress balance compensation and suppression point in the typical thin-walled tooth sensitive region, the multiple stress balance compensation and suppression points in the remaining thin-walled tooth region and the reverse pulse pressure of each stress balance compensation and suppression point in the remaining thin-walled tooth region are determined.

8. The intelligent optimization system for aerospace component processing parameters as described in claim 5, characterized in that, The stress clustering module is also used for: The residual stress detection data from the multiple basic residual stress detection points were clustered using the K-means clustering algorithm to obtain multiple residual stress distribution clusters.

9. An electronic device, characterized in that, include: processor; Memory; And a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the intelligent optimization method for the processing parameters of aerospace parts as described in any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the intelligent optimization method for the processing parameters of aerospace parts as described in any one of claims 1 to 4.

Images