A simulated grid crack repair method and aircraft engine shape optimization method

Abstract

The present invention discloses a method for repairing fine cracks in a simulation grid and a method for optimizing the appearance of an aircraft engine. The method divides fine cracks in a simulation grid into pseudo fine cracks, connected fine cracks, and non-connected fine cracks. Pseudo fine cracks are eliminated by global vertex merging. Complex structure holes are separated by local vertex merging. Connected fine cracks are screened out in holes based on aspect ratio and area. Non-connected fine cracks are screened out based on spatial position and shape characteristics. The optimal suture points for connected fine cracks and non-connected fine cracks are then screened. The present invention achieves fast and robust fine crack repair and can locally repair fine cracks in a triangular grid without destroying the original grid geometry. Since the present invention does not destroy the original geometry of the model, it is suitable for simulation pre-processing of high-precision models such as aircraft engines, thereby achieving optimized design of aircraft engine appearance.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial numerical simulation, and in particular relates to a simulation grid fine crack repair method and an aircraft engine shape optimization method. Background Art

[0002] Discrete models represented by triangular meshes are widely used in engineering applications, such as product design and manufacturing. In high-tech fields such as aircraft manufacturing, it is necessary to first obtain the meshes of aircraft components such as wings and engines through modeling, and then perform simulation calculations. Only when the simulation results meet expectations can production experiments be carried out. However, the triangular mesh of the aircraft engine is mainly discretized from the CAD model. Due to the extreme complexity of the engine, the obtained triangular mesh often contains defects such as cracks, holes, self-intersections, non-manifold elements and inconsistent surface orientations. Among these defects, gaps and holes are considered to be the most important defects, which will cause the failure of many downstream algorithms, such as constructing body meshes and simulation algorithms. Existing crack repair methods are only based on geometric information and cannot robustly sew all cracks. Therefore, how to quickly and robustly repair cracks so that the aircraft engine model can correctly perform engineering simulations has become an urgent problem to be solved. Summary of the Invention

[0003] To address the engineering issues of aircraft engine simulation pre-processing, the present invention provides a simulation mesh crack repair method and an aircraft engine shape optimization method. The method can identify different types of cracks in the triangular mesh of the aircraft engine model and perform feature-preserving stitching to repair the cracks in the mesh, enabling downstream simulation work.

[0004] In a first aspect, a method for repairing a simulated grid crack comprises the following steps:

[0005] Step 1: Set three types of cracks: pseudo cracks, connected cracks, and disconnected cracks. Pseudo cracks are formed by points whose distance is less than a global threshold; disconnected cracks are formed by matching two adjacent holes; and connected cracks are connected cracks formed by a single hole and are not pseudo cracks. For the entire triangle mesh, perform boundary vertex merging based on a global threshold to eliminate pseudo cracks. Extract holes from the triangle mesh; for each hole, perform boundary vertex merging based on a dynamically adjusted local threshold, separating complex holes into multiple simpler ones.

[0006] Step 2: Based on aspect ratio and area, connective slits are screened out from the holes; based on spatial position and shape characteristics, non-connective slits are screened out;

[0007] Step 3, suturing of slits: For connected slits, the optimal suturing point is found for each boundary vertex among the boundary edges or vertices of the same connected slit; for non-connected slits, the optimal suturing point is found for each boundary vertex of one hole among the boundary edges or vertices of the other hole.

[0008] Preferably, in step 1, each hole corresponds to a local threshold; the local threshold is 1 / 5 to 1 / 3 of the length of the shortest boundary edge in the hole.

[0009] Preferably, the process of screening out connected slits in step 2 is as follows:

[0010] (1) For all holes processed in step 1, if two boundary edges meet the following conditions: there is a common vertex, and the common vertex is adjacent to only two boundary edges, then merge the union-find sets to which the two boundary edges belong to obtain multiple subsets. Collect simple holes in the subsets. The simple holes are holes in which all boundary vertices are adjacent to only two boundary edges.

[0011] (2) Determine whether each simple hole is a connected crack: For a simple hole with more than two sharp points, if the aspect ratio of the simple hole is greater than the first threshold, the simple hole is determined to be a connected crack. For a simple hole with no sharp point or only one sharp point, if the area S of the simple hole is greater than the first threshold, the simple hole is determined to be a connected crack. H and the average area S of the 1-neighborhood patches avg If the ratio is less than the second threshold, the simple hole is determined to be a connected crack.

[0012] Preferably, the process of collecting simple holes in subsets is as follows: determining whether each subset is a simple hole one by one; matching subsets that are not determined to be simple holes; if two subsets have the same boundary vertices with a boundary degree of 1, then unioning the two subsets to form a simple hole; and deleting the remaining subsets. The boundary degree is the number of boundary edges connected by the boundary vertex in the current subset.

[0013] Preferably, the first threshold value ranges from 4.0 to 5.0; the second threshold value ranges from 0.2 to 0.3.

[0014] As a preferred method, the process of screening out non-connected slits in step 2 is as follows: for any two holes 、 , calculate the Hausdorff distance; if the hole With holes The Hausdorff distance between the holes is smaller than The average length of the boundary edge is , then mark the hole Holes Candidate holes that constitute non-connected cracks; holes Each candidate hole is in the hole Calculate the area similarity; select the candidate hole with the highest area similarity as the best matching hole; hole It forms a non-connected slit with the optimal matching hole.

[0015] Preferably, in step 3, the process of finding the optimal suture point is as follows:

[0016] (1) For boundary vertices , take the threshold The boundary edges within the range are used as the set to be matched . Traverse the set to be matched Each edge to be matched in ; If the boundary vertex Falling on the side to be matched If the boundary vertex is within the vertical domain of To the edge to be matched The projection distance of the edge to be matched is the shortest. Insert a virtual point on the boundary vertex Form a VR point pair. If the position of the virtual point is close to an edge to be matched The boundary vertex on , then use the edge to be matched The boundary vertices on the Form an rr point pair;

[0017] (2) First, remove duplicate rr point pairs. Second, if a boundary vertex constitutes both an rr point pair and a vr point pair, delete the vr point pair corresponding to the boundary vertex and the corresponding virtual point at the same time.

[0018] (3) Convert all remaining virtual points into new boundary vertices.

[0019] (4) Perform point-pair intersection detection and swap vertices for intersecting point pairs to eliminate the intersection.

[0020] (5) Perform point-to-point merging to complete fine seam suture.

[0021] Preferably, the threshold Set to the average side length of the slits.

[0022] As a preference, the original boundary vertex is used as the priority point; for a point pair with only one priority point, the priority point is used as the merged boundary vertex; for the remaining point pairs, the midpoint of the two boundary vertices is used as the merged boundary vertex.

[0023] In a second aspect, the present invention provides a method for optimizing the shape of an aircraft engine, comprising the following steps:

[0024] Step 1: Divide the surface of the aircraft engine 3D geometric model into triangular meshes.

[0025] Step 2: Repair the fine cracks in the triangular mesh obtained by step 1 by using the fine crack repair method of the simulation mesh as claimed in claim 1.

[0026] Step 3: Generate a volume mesh based on the repaired triangular mesh.

[0027] Step 4: Perform fluid dynamics simulation based on the volume mesh; optimize the shape of the aircraft engine's three-dimensional geometric model based on the simulation results.

[0028] Step 5: Repeat steps 1 to 4 until the fluid mechanics simulation results meet the design requirements.

[0029] The present invention has the following beneficial effects.

[0030] 1. This invention classifies thin seams into three types: pseudo-slits, connected thin seams, and non-connected thin seams, and provides targeted suturing. For non-connected thin seams formed by two adjacent holes, the suture points for the vertices of each hole are found in the other hole, effectively suturing non-connected thin seams with little impact on the overall grid structure. Furthermore, this invention designs a thin seam suturing method based on feature preservation using prioritized points.

[0031] 2. Based on the topological and geometric characteristics of connected and non-connected cracks, the present invention proposes a connected crack detection method based on topology, aspect ratio, and area, and a hole matching method based on spatial position and geometric shape to identify non-connected cracks, thereby improving the accuracy and comprehensiveness of crack identification.

[0032] 3. This invention implements fast and robust crack repair technology, capable of locally repairing cracks in triangular meshes without destroying the original mesh geometry. Because this method preserves the model's geometry, it is suitable for simulation pre-processing of high-precision models such as aircraft engines, enabling optimized design of aircraft engine shapes. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 Schematic diagram of different types of slits in step 0 of an embodiment of the present invention.

[0034] Figure 2 Schematic diagram of the global-local vertex merging result in step 1 of an embodiment of the present invention.

[0035] Figure 3Schematic diagram of simple hole segmentation in step 2 of an embodiment of the present invention.

[0036] Figure 4 Schematic diagram of hole width calculation in step 2 of an embodiment of the present invention.

[0037] Figure 5 Schematic diagram of the vertical domain of the boundary edge in step 3 of an embodiment of the present invention.

[0038] Figure 6 Schematic diagram of vr vertex pairs and rr vertex pairs in step 3 of an embodiment of the present invention.

[0039] Figure 7 Schematic diagram of the crossover detection and exchange of rr vertex pairs in step 3 of an embodiment of the present invention.

[0040] Figure 8 2. This is a comparison diagram before and after mesh suture according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be further described below with reference to the accompanying drawings.

[0042] A method for repairing fine cracks in a simulation mesh. In this embodiment, the simulation mesh is specifically a triangular mesh of a three-dimensional geometric model of an aircraft engine. This embodiment is used to identify different types of fine cracks in the triangular mesh of the aircraft engine model and perform feature-preserving stitching to repair the fine cracks in the mesh, enabling downstream simulation work and solving engineering problems related to aircraft engine simulation pre-processing.

[0043] In this embodiment, the slits in the triangular mesh are classified in detail and classified into three types: pseudo slits, connected slits and non-connected slits. Subsequently, global-local vertex merging is performed to repair pseudo slits: a global threshold is used to merge all similar boundary vertices, and then for each hole, an adaptive local threshold is calculated, and the boundary vertices in the hole are merged according to the threshold. Secondly, slit identification is performed: all simple holes are checked to determine whether they are connected slits or non-connected slits. Finally, for each slit, a feature-preserving stitching method is used to stitch the slits. Finally, for all stitched edges, edge optimization is performed, and the shorter edges are collapsed to improve the quality of the mesh. Specifically, the following steps are included:

[0044] Step 0. The present invention classifies slits in the aircraft engine simulation grid into three types: pseudo slits, connected slits, and non-connected slits. Pseudo slits are caused by repeated vertices with the same coordinates. Connected slits are simple, narrow holes. The boundary vertices of these slits are connected, and there are generally at least two very small angles formed by adjacent edges. Non-connected slits are complex and special slits. These slits are formed by matching two holes of similar shape and close distance. They are connected by no or only one non-manifold vertex. Figure 1 Three different types of fine cracks are shown, where (a) is a pseudo fine crack, (b) is a connected fine crack, and (c) is a non-connected fine crack.

[0045] Step 1: Perform global-local vertex merging. First, use a smaller global threshold to merge all boundary vertices on the triangle mesh. All similar boundary vertices within this global threshold will be merged into one vertex. This threshold is generally set by the user or directly set to the length of the shortest edge in the mesh. In this embodiment, the global threshold is set to 10 -5 This step can eliminate the existence of pseudo-slits.

[0046] Then, for each simple hole in the simulation mesh , calculate a local threshold Used to merge simple holes whose distance is less than the local threshold To prevent the original edge from collapsing due to merging, this embodiment calculates the local threshold based on the shortest edge length of the boundary edge in the hole. :

[0047] (1)

[0048] in, For simple holes 1 / 4 is the most preferred coefficient in this embodiment; in some other embodiments, the coefficient may also take other values ​​less than 1, such as 1 / 3 or 1 / 5.

[0049] Merging the vertices of each simple hole according to the threshold can improve the stitching rate of fine seams and simplify the shape of complex holes, that is, dividing a complex hole containing narrow and long areas and non-narrow and long areas into multiple independent holes with simple shapes.

[0050] Figure 2 Demonstrates the use of local thresholding The result of this process of merging boundary vertices is Figure 2 Part (a) is the original mesh with fine cracks. Figure 2 Part (b) is the result of global merging. Figure 2Part (c) is the result of local merging, where a hole with a complex shape is simplified into two holes with simple shapes.

[0051] Step 2: Different types of cracks in the aircraft engine mentioned in Step 0 need to be detected and identified. This embodiment proposes a method for detecting connected cracks based on aspect ratio and area, and a method for detecting non-connected cracks based on spatial position and shape features:

[0052] Step 2-1. Find all simple holes. A simple hole is defined as a hole in which all boundary vertices are adjacent to only two boundary edges. This process is implemented by union-find. The hole segmentation algorithm using union-find is as follows:

[0053] (1) Use union-find to split the input connected boundary edge set into multiple subsets. For each boundary edge, traverse all other boundary edges if they meet the following conditions: and If there are only two adjacent boundary edges, the union-find sets to which the two boundary edges belong are merged to obtain multiple subsets. According to the definition of simple holes, determine whether each subset is a simple hole and collect all simple holes; if there are still subsets that have not been processed, go to step (2).

[0054] (2) Match the remaining boundary edge subsets (a boundary edge subset must contain two boundary vertices with a boundary degree of 1 relative to the subset); the boundary degree is the number of boundary edges connected to the boundary vertex in the current subset. If the vertices with a boundary degree of 1 in the two subsets are the same, then the two subsets are unioned to form a simple hole; otherwise, the subsets are discarded.

[0055] Figure 3 The following is a flowchart for hole segmentation; (a) is the original mesh, where the red points are boundary vertices and the dark blue edges are boundary edges. According to the hole detection algorithm, they will be divided into two connected boundary edge sets, one of which contains boundary edges on non-manifold patches. (b) is the result of using union-find to perform a union operation on each boundary edge set. Edges in the same union-find set have the same color, and the blue, purple, and green boundary edge subsets are all judged and extracted as simple holes. (c) is the result of endpoint matching on the remaining unmatched boundary edge subsets. The original red and orange subsets are matched to form a boundary edge subset, while the gray subset is not matched and will be discarded in the next stage. (d) is the final segmentation result, with a total of 4 simple holes segmented.

[0056] Step 2-2. After collecting all simple holes, determine whether they are connected cracks: The overall process of connected crack detection is as follows:

[0057] (1) Identify cusps on the hole (boundary vertices with adjacent boundary angles less than 30°). Each pair of adjacent cusps forms a group. If the number of cusps is less than 2, proceed to step (4). In this embodiment, the angle threshold for identifying cusps is 30°. In other embodiments, the angle threshold can also be other values, such as 20°, 25°, 35°, etc.

[0058] (2) Calculate the maximum value of the shortest distance from the boundary vertex in each group to the boundary edge in other groups. This maximum value is used to describe the width of the hole. The total length of the boundary edge in the hole is calculated to calculate the aspect ratio.

[0059] (3) If the aspect ratio is > 4, the hole is identified as a connected slit.

[0060] (4) If the number of cusps is less than 2, triangulate the hole and calculate the hole area S H , calculate the average area S of the 1-neighborhood patches of the hole avg ; If the area of ​​the hole S H < 0.2·S avg , the hole is identified as a connected slit.

[0061] Figure 4 Figure 1 shows the flow chart for hole width calculation. (a) shows a simple hole after hole detection and segmentation. (b) shows the result after cusp detection, where red vertices are cusps. (c) shows the result after grouping, where non-cusp vertices and boundary edges are divided into three groups. (d) shows the result of width calculation. For each non-cusp vertex, a shortest projection distance is calculated, and the maximum of these shortest projection distances is used as the final width W.

[0062] Step 2-3. For the holes that are not judged as connected cracks, continue to judge whether they can form non-connected cracks with other holes. , find other similar holes in its space , then calculate the Hausdorff distance of the two hole point sets. For two holes, if the Hausdorff distance is small, they will have a greater probability of matching into a non-connected crack. With holes The Hausdorff distance between the holes is smaller than The average length of the boundary edge is , then mark the hole Holes Candidate holes that constitute non-connected cracks, for holes There may be multiple candidate holes that can form a non-connected crack. Therefore, it is necessary to select an optimal candidate hole and hole Matching becomes a non-connected slit.

[0063] In order to quantify the optimal index, this paper proposes the area similarity measurement of two holes. Holes The area of ​​the hole and holes The area similarity between The measurements are:

[0064] (2)

[0065] From the hole Select a hole with the largest similarity measure from all candidate holes As the optimal match, they are combined into a non-connected slit.

[0066] Step 3: To close the slit, find an optimal suturing point for all non-sharp points in the slit, and after determining all the suturing point pairs, perform a merge operation to close the slit. The algorithm process includes:

[0067] Step 3-1. Matching the optimal stitching point of the boundary vertex. For each vertex in the slit, match an optimal stitching point. The stitching point can be a projection point on the edge or another boundary vertex (if the slit is a non-connected slit, the matching stitching point cannot come from the hole where the vertex is located). The specific matching process for the optimal stitching point of the boundary vertex is as follows:

[0068] First, each boundary vertex is queried within the threshold All the boundary edges of the slit within the range are used as the set to be matched (If it is a non-connected crack, remove the boundary edge of the hole where the current vertex is located.) Threshold is set to the average side length of the slits, and the threshold If it is set larger, some fine cracks that are closer to the shape of holes can be stitched together, but it may cause the mesh geometry to be destroyed. The set to be matched Then, traverse the set to be matched Each edge to be matched in , if the edge to be matched is a boundary vertex If the adjacent edge is , skip it; otherwise, judge the boundary vertex Whether it falls on the edge to be matched within the vertical domain. Figure 5 Shows the vertical domain of the boundary edge. If the boundary vertex Falling on the side to be matched If it is outside the vertical domain of Find the suture point on the boundary vertex Falling on the side to be matched If the boundary vertex is within the vertical domain of To the edge to be matched Projection distance After traversing all the edges to be matched Then, on the edge to be matched with the shortest projection distance Insert a virtual point on the vertex Form a vr point pair (virtual-real). The virtual point is the boundary vertex To the edge to be matched If the virtual point is very close to any vertex on this edge (i.e., the distance is less than 1% of the edge length), we can directly construct a rr point pair (real-real) between the two vertices and discard the virtual point.

[0069] Figure 6 The result is shown, where the red vertices are sharp points, the blue vertices are vertices that exist in the slit, and the pink vertices are virtual points.

[0070] Step 3-2. Point pair optimization. This step is used to remove duplicate matching vertices. The specific process of point pair optimization is as follows:

[0071] In step 3-1, the initial vr point pairs and rr point pairs are obtained. This step optimizes these point pairs, specifically point pair deduplication. Point pair deduplication is necessary because duplicate rr point pairs may be generated during the matching of boundary vertices, and it is possible that a boundary vertex participates in both the rr point pair and the vr point pair, which is not allowed. Therefore, this embodiment deduplicates all point pairs according to the following rules, first removing duplicate rr point pairs. Secondly, if a boundary vertex constitutes both an rr point pair and a vr point pair, delete the vr point pair corresponding to the boundary vertex, and delete the corresponding virtual point at the same time.

[0072] Step 3-3. Split the face. This step converts all virtual points into real vertices. The triangle face adjacent to the boundary edge will be split into multiple sub-triangles based on the virtual vertices on the edge. The specific process of face splitting is as follows:

[0073] For all virtual points inserted on the boundary edge, sort them in ascending order based on their distance from the starting point of the boundary edge. Then perform a patch splitting operation so that all virtual points become real vertices and all vr point pairs are converted to rr point pairs.

[0074] Step 3-4. Point-pair intersection detection and exchange. This step is used to predict whether the topology after stitching is correct and to prevent the generation of non-manifolds. The specific process of point-pair intersection detection and exchange is as follows:

[0075] For a certain rr point pair The last two matching vertices and , randomly select one of the vertices and check the adjacent vertices of the vertex in the slit If the adjacent vertices With matching rr point pairs , then check the rr point pair Are the two vertices on the rr point pair? If they are on the same side, no intersection occurs; otherwise, the rr point pair Point pair with rr If an intersection occurs, swap the two pairs of points to form two new pairs of non-intersecting rr points. Loop through all boundary vertices and perform checks and swaps until no intersection occurs.

[0076] Step 3-5. Point-to-point merging. The specific process is as follows:

[0077] The last operation of stitching the slits is to merge all the rr point pairs. The present invention adopts a merging method based on priority points. There are two sources of vertices in the rr point pairs, one is the boundary vertices in the original slits, and the other is converted from the virtual vertices in the vr point pairs. The boundary vertices in all the original slits are marked as priority points. Therefore, there are three possible compositions of the rr point pairs: 1. There is only one priority point, and the original boundary vertices and virtual vertices are combined. 2. There are two priority points, and the original boundary vertices and the original boundaries are combined. 3. There are no priority points, and the virtual vertices and the virtual vertices are combined (caused by exchange after intersection). For the first combination, the coordinates of the merged point are set to the position of the priority point; for the second and third combinations, the coordinates of the merged point are set to the midpoint position of the two vertices. This method follows the information of the original mesh and can ensure that the features are as close to the original mesh as possible.

[0078] Figure 7 Shows the process, Figure 7 Part (a) shows all the rr point pairs after the face is split. Figure 7 Part (b) of the algorithm detects pairs of vertices that intersect. Figure 7 Part (c) is the result after exchanging vertices. Figure 7 Part (d) is the result of fine seam suture.

[0079] Step 4: Optimize all stitched edges. If the length of the stitched edge is less than 0.15 times the average length of the original slit, perform edge collapse to improve the quality of the mesh.

[0080] The suture method provided in this embodiment is used to repair the mesh cracks on the three-dimensional model. The comparison before and after the repair is shown as follows: Figure 8 As shown in the figure, the pink edges are the boundary edges in the original mesh. Figure 8 It can be seen that the four models become watertight models after repair and retain the geometric characteristics of the original mesh.

[0081] This embodiment can effectively repair fine cracks in aircraft engines or other CAD models, and preserve the geometry and topology of meshes in other parts, thus supporting subsequent simulation work on aircraft engines.

[0082] Example 2

[0083] A method for optimizing the shape of an aircraft engine comprises the following steps:

[0084] Step 1: Divide the surface of the aircraft engine 3D geometric model into triangular meshes.

[0085] Step 2: Repair the fine cracks in the triangular mesh obtained by step 1 by using the fine crack repair method of the simulation mesh provided in Example 1.

[0086] Step 3: Generate a volume mesh based on the repaired triangular mesh.

[0087] Step 4: Perform fluid dynamics simulation based on the volume mesh; optimize the shape of the aircraft engine's three-dimensional geometric model based on the simulation results.

[0088] Step 5: Repeat steps 1 to 4 until the fluid mechanics simulation results meet the design requirements.

Claims

1. A simulated grid crack repair method: characterized by: The following steps are involved: Step 1: For cracks in the triangular mesh of the 3D geometric model of the aircraft engine, three crack types are set: pseudo cracks, connected cracks, and non-connected cracks. Pseudo cracks are formed by points whose distance is less than a global threshold; non-connected cracks are formed by matching two adjacent holes; and connected cracks are formed by a single hole and are not pseudo cracks. For the entire triangle mesh, the boundary vertices are merged based on the global threshold to eliminate pseudo-slits; for each hole, the boundary vertices are merged based on the dynamically adjusted local threshold; Step 2: Based on aspect ratio and area, connective slits are screened out from the holes; based on spatial position and shape characteristics, non-connective slits are screened out; Step 3, suturing of slits: For connected slits, the optimal suturing point is found for each boundary vertex among the boundary edges or vertices of the same connected slit; for non-connected slits, the optimal suturing point is found for each boundary vertex of one hole among the boundary edges or vertices of the other hole.

2. The method for repairing simulated mesh cracks according to claim 1, characterized in that: In step 1, each hole corresponds to a local threshold; the local threshold is 1 / 5 to 1 / 3 of the length of the shortest boundary edge in the hole.

3. The method for repairing simulated mesh cracks according to claim 1, characterized in that: The process of screening out connected slits in step 2 is as follows: (1) For all holes processed in step 1, if two boundary edges meet the following conditions: there is a common vertex, and the common vertex is adjacent to only two boundary edges; then the union-find set to which the two boundary edges belong is merged to obtain multiple subsets; simple holes are collected in the subsets; the simple holes are holes in which all boundary vertices are adjacent to only two boundary edges; (2) Determine whether each simple hole is a connected crack: For a simple hole with more than two sharp points, if the aspect ratio of the simple hole is greater than the first threshold, the simple hole is determined to be a connected crack; for a simple hole with no sharp point or only one sharp point, if the area S of the simple hole is greater than the first threshold, the simple hole is determined to be a connected crack. H and the average area S of the 1-neighborhood patches avg If the ratio is less than the second threshold, the simple hole is determined to be a connected crack.

4. The method for repairing simulated mesh cracks according to claim 3, characterized in that: The process of collecting simple holes in subsets is as follows: determine whether each subset is a simple hole one by one; match the subsets that are not determined to be simple holes; if the boundary vertices with a boundary degree of 1 in two subsets are the same, then the two subsets are unioned to form a simple hole; and the remaining subsets are deleted.

5. The method for repairing simulated mesh cracks according to claim 3, characterized in that: The first threshold value is between 4.0 and 5.0; the second threshold value is between 0.2 and 0.

3.

6. The method for repairing simulated mesh cracks according to claim 1, characterized in that: The process of screening out non-connected slits in step 2 is as follows: for any two holes 、 , calculate the Hausdorff distance; if the hole With holes The Hausdorff distance between the holes is smaller than The average length of the boundary edge is , then mark the hole Holes Candidate holes that constitute non-connected cracks; holes Each candidate hole is in the hole Calculate the area similarity; select the candidate hole with the highest area similarity as the best matching hole; hole It forms a non-connected slit with the optimal matching hole.

7. The method for repairing simulated mesh cracks according to claim 1, characterized in that: In step 3, the process of finding the optimal stitching point is as follows: (1) For boundary vertices , take the threshold The boundary edges within the range are used as the set to be matched ; Traverse the set to be matched Each edge to be matched in ; If the boundary vertex Falling on the side to be matched If the boundary vertex is within the vertical domain of To the edge to be matched The projection distance of the edge to be matched is the shortest. Insert a virtual point on the boundary vertex Form a VR point pair; if the position of the virtual point is close to an edge to be matched The boundary vertex on , then use the edge to be matched The boundary vertices on the Form an rr point pair; (2) First, remove duplicate rr point pairs; second, if a boundary vertex constitutes both an rr point pair and a vr point pair, delete the vr point pair corresponding to the boundary vertex and the corresponding virtual point at the same time; (3) Convert all remaining virtual points into newly added boundary vertices; (4) Perform point-pair intersection detection and perform vertex swapping for intersecting point pairs; (5) Perform point-to-point merging to complete fine seam suture.

8. The method for repairing simulated mesh cracks according to claim 7, characterized in that: Threshold Set to the average side length of the slits.

9. The method for repairing simulated mesh cracks according to claim 7, characterized in that: The original boundary vertex is taken as the priority point; for point pairs with only one priority point, the priority point is taken as the merged boundary vertex; for the remaining point pairs, the midpoint of the two boundary vertices is taken as the merged boundary vertex.

10. A method for optimizing the shape of an aircraft engine, characterized by: The following steps are involved: Step 1: Perform triangular meshing on the surface of the three-dimensional geometric model of the aircraft engine; Step 2: Repairing the fine cracks in the triangular mesh obtained by step 1 by using the fine crack repair method of the simulation mesh as claimed in claim 1; Step 3: Generate a volume mesh based on the repaired triangular mesh; Step 4: Perform fluid dynamics simulation based on the volume mesh; optimize the shape of the aircraft engine's three-dimensional geometric model based on the simulation results; Step 5: Repeat steps 1 to 4 until the fluid mechanics simulation results meet the design requirements.

Images