A pyrolytic carbon interfacial phase modeling and debonding behavior analysis method based on graph theory

By employing graph theory-based methods and the moving window method, the simulation problem of global crack propagation path in the pyrolytic carbon interface phase was solved, achieving efficient and accurate crack path calculation and parameter extraction for the PyC interface phase, and supporting the optimized design of ceramic matrix composites.

CN123120125APending Publication Date: 2026-11-17BEIHANG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610914483.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-06-24
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing theories on interfacial debonding have failed to effectively address the global propagation path of cracks within the pyrolytic carbon interfacial phase. Traditional methods are computationally expensive and struggle to handle the nonlinear mechanical behavior resulting from a large number of anisotropic subdomains.

Method used

A graph-based approach is used to discretize the pyrolytic carbon interface phase into polygonal mesh elements, assign fracture energy to the mesh edges, identify crack propagation paths by global energy minimization, and dynamically update the local computational domain using the moving window method to simulate crack propagation within the PyC interface phase.

Benefits of technology

It achieves efficient solution for intraphase crack propagation at PyC interfaces, obtains the global energy minimization path, can describe long-distance crack propagation, extract key interface parameters, and supports interface optimization design of ceramic matrix composites.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120125A_ABST
    Figure CN123120125A_ABST
Patent Text Reader

Abstract

The application discloses a pyrolytic carbon interface phase modeling and debonding behavior analysis method based on graph theory. The method comprises the following steps: obtaining a SiC-restrained PyC interface phase model; assigning a fracture energy to all grid edges of the obtained SiC-restrained PyC interface phase model, thereby obtaining a PyC grid model in which all grid edges are assigned with corresponding fracture energies; calculating a crack propagation path in a first window according to the PyC grid model in which all grid edges are assigned with corresponding fracture energies; obtaining a global crack propagation path in a PyC interface phase according to the crack propagation path in the first window and the PyC grid model in which all grid edges are assigned with corresponding fracture energies; and calculating key interface parameters according to the global crack propagation path in the PyC interface phase. The application can establish a graph model with a large number of different orientation subdomain materials, and is particularly suitable for materials such as PyC with a significant layered subdomain structure.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of failure analysis technology for pyrolytic carbon interface phases, specifically to a graph theory-based method for modeling pyrolytic carbon interface phases and analyzing their debonding behavior. Background Technology

[0002] Ceramic matrix composites possess excellent properties such as high specific strength and high temperature resistance, making them ideal materials for hot-end components of high thrust-to-weight ratio aero-engines. Pyrolytic carbon (PyC) is a commonly used interfacial phase in ceramic matrix composites, and its debonding behavior has a crucial impact on the macroscopic mechanical properties of ceramic matrix composites by influencing interfacial fracture energy and interfacial roughness. The debonding behavior of PyC is controlled by its microstructure. Depending on the preparation process, PyC exhibits different texture characteristics. In experiments, the orientation angle OA is used as a parameter to quantitatively describe the dispersion of the orientation of graphite subdomains within PyC. OA, from small to large, corresponds to high, medium, and low texture PyC, respectively, meaning that the dispersion of internal subdomain orientation gradually increases. The complex microstructure of PyC poses a challenge to the study of its debonding behavior.

[0003] Existing theories on interfacial debonding primarily focus on whether cracks deflect or penetrate at the interface, neglecting the continuous crack propagation path within the interfacial phase. In fact, due to the large number of anisotropic subdomains within the PyC interfacial phase, traditional fracture mechanics methods struggle to solve for crack propagation behavior within PyC. Furthermore, common numerical simulation methods such as the phase-field method also have limitations in this type of problem: the computational cost is extremely high due to the need to handle the nonlinear mechanical behavior arising from numerous anisotropic subdomains, and the phase-field method's local energy minimization solution may not yield the globally optimal crack path.

[0004] In recent years, scholars have proposed graph theory-based crack propagation simulation methods. These methods discretize the material into polygonal mesh elements and embed the material's microscopic fracture energy information into the mesh edges. Finally, based on the principle of global energy minimization, the graph cut path that minimizes energy is identified, which is the crack propagation path. This method has unique advantages for crack propagation problems dominated by anisotropic fracture energy, as it is relatively easy to model and computationally efficient. However, current graph theory-based crack propagation simulation research only focuses on single-pass crack deflection problems within a few dozen grains. Further methodological improvements and developments are needed for crack propagation problems within PyC interface phases containing tens of thousands of subdomains and with elongated morphologies. Summary of the Invention

[0005] The purpose of this invention is to provide a graph theory-based method for modeling pyrolytic carbon interface phases and analyzing debonding behavior to at least solve one of the aforementioned technical problems.

[0006] One aspect of the present invention provides a graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interface phases, the graph-based method comprising:

[0007] Step 1: Obtain the PyC interface phase model with SiC constraints;

[0008] Step 2: Assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy.

[0009] Step 3: Obtain the global crack propagation path within the PyC interface phase based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy;

[0010] Step 4: Calculate the key interface parameters based on the global crack propagation path within the PyC interface phase.

[0011] Optionally, step 3 includes:

[0012] Step 31: Calculate the crack propagation path within the first window based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy;

[0013] Step 32: Obtain the global crack propagation path within the PyC interface phase based on the crack propagation path within the first window and the PyC mesh model where all mesh edges are assigned corresponding fracture energies.

[0014] Optionally, step 1 includes:

[0015] Step 11: Obtain the complete set of basic feature parameters for the PyC interface phase to be modeled;

[0016] Step 12: Define the geometric region of the PyC interface phase based on the complete set of basic feature parameters of the PyC interface phase to be modeled, thereby obtaining the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries.

[0017] Step 13: Generate a set of N parameters based on the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries and the complete set of basic feature parameters of the PyC interface phase to be modeled. s PyC subdomain topology of a Voronoi polygon;

[0018] Step 14: Generate an acute-angled triangular mesh set covering the entire PyC computational domain based on the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries and the complete set of basic feature parameters of the PyC interface phase to be modeled.

[0019] Step 15: Identify the subdomain to which each mesh belongs based on the PyC subdomain topology and the acute-angled triangle mesh set, thereby obtaining a set of triangle meshes with subdomain number labels;

[0020] Step 16: Perform smooth subdomain boundary optimization based on the triangular mesh set with subdomain number labels to obtain a PyC mesh model with smooth continuous subdomain boundaries;

[0021] Step 17: Generate a unique weak surface direction for each subdomain based on the complete set of fundamental feature parameters of the PyC mesh model with smooth continuous subdomain boundaries and the PyC interface phase to be modeled. The PyC mesh model, wherein each subdomain has a unique weak surface direction. The PyC mesh model is used as the discrete mesh model of the PyC interface phase;

[0022] Step 18: Introduce SiC constraints on both sides of the PyC interface phase discrete mesh model to obtain a PyC interface phase model with SiC constraints.

[0023] Optionally, step 2 includes:

[0024] Step 21: Extract all mesh edges of the PyC interface phase model with SiC constraints and classify them to obtain the set of classified mesh edges;

[0025] Step 22: Calculate the fracture energy of the mesh edges in the subdomain based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all mesh edges in the subdomain;

[0026] Step 23: Calculate the fracture energy of the inter-subdomain mesh edges based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all inter-subdomain mesh edges;

[0027] The fracture energy values ​​corresponding to the mesh edges within all subdomains and the fracture energy values ​​corresponding to the mesh edges between all subdomains are assigned fracture energy attributes to the classified mesh edge set, thereby obtaining a PyC interface phase model with SiC constraints in which all mesh edges are assigned corresponding fracture energy attributes.

[0028] Optionally, step 31 includes:

[0029] Extract the first computational window of the PyC mesh model where all mesh edges are assigned the corresponding fracture energy, thereby obtaining the independent computational domain of the first window;

[0030] Initial crack boundary conditions and external load conditions are set for the independent computational domain of the first window, thereby obtaining the computational domain of the first window with fixed initial crack boundary conditions.

[0031] The crack tip stress intensity factor is calculated based on the first window computational domain with fixed initial crack boundary conditions and external load conditions, thereby obtaining the crack tip type I stress intensity factor and crack tip type II stress intensity factor.

[0032] The energy release rate as a function of the crack deflection angle is obtained based on the type I and type II stress intensity factors at the crack tip.

[0033] The set of cost function values ​​for all mesh elements within the first window is generated based on the energy release rate as a function of the crack deflection angle and the first window computational domain with fixed initial crack boundary conditions.

[0034] Obtain the total cost function based on the set of cost function values ​​of all grid cells in the first window;

[0035] Based on the total cost function and the first window computational domain with fixed initial crack boundary conditions, a constrained graph cut optimization is performed to find the minimum cost crack path, thereby obtaining the initial crack propagation path within the first window.

[0036] Quasi-static propagation control is applied to the initial crack propagation path within the first window to obtain the complete crack propagation path within the first window that satisfies the quasi-static propagation conditions.

[0037] Optionally, step 32 includes:

[0038] Step 321: From the complete crack propagation path within the first window that satisfies the quasi-static propagation condition, taking the initial crack tip as the starting point, cut off the length after the starting point as the preset length. The crack segment is added to the global crack propagation path; a preset length is then cut off from the tail end of the crack segment along the crack propagation direction. The crack segment is used as the initial crack in the second calculation window, and the inclination angle of the initial crack in the second window is calculated.

[0039] Step 322: Translate the first calculation window along the direction of macroscopic crack propagation until the window boundary coincides with the initial crack initiation point of the second calculation window, thus creating the independent calculation domain of the second calculation window;

[0040] Step 323: Within the independent computational domain of the second computational window, solve for the complete crack propagation path that satisfies the quasi-static propagation condition within the second window using the same computational method as the first window;

[0041] Step 324: Repeat the following operation until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. :

[0042] Step 325: From the complete crack propagation path of the currently solved k-th window, extract the length after the crack tip of the initial crack in the k-th window, which is... The crack segment is added to the global crack propagation path; where, ;

[0043] Step 326: Cut a length of [length missing] from the tail end of the crack segment in the opposite direction. The crack segment is used as the initial crack in the (k+1)th calculation window, and its dip angle is calculated.

[0044] Step 327: Following the methods in steps 322 to 323, generate the independent computational domain of the (k+1)th computational window and solve for the complete crack propagation path that satisfies the quasi-static propagation condition;

[0045] Step 328: Continue until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. When the last calculation window is obtained, the complete crack propagation path of the last window is obtained by solving according to the methods in steps 322 to 323;

[0046] Wherein, if the length obtained during the extraction of the crack segment in any window is... If the absolute value of the initial crack inclination angle is greater than or equal to 90°, then the crack extension segment entering the global crack propagation path is lengthened, and the tail length is re-trunculated. The crack segment is used as the initial crack for the next window until the absolute value of the initial crack dip angle is less than 90°.

[0047] Step 329: Piece together all the crack segments in the global crack propagation path according to the window progression order to obtain the global crack propagation path within the PyC interface.

[0048] Optionally, step 4 includes:

[0049] Step 41: Calculate the interfacial fracture energy based on the global crack propagation path within the PyC interface phase, thereby obtaining the interfacial fracture energy of the PyC interface phase.

[0050] Step 42: Calculate the interface roughness based on the global crack propagation path within the PyC interface phase, thereby obtaining the interface roughness of the PyC interface phase.

[0051] Optionally, step 22 includes:

[0052] The fracture energy of each mesh edge within a subdomain is calculated using the following formula:

[0053] ;

[0054] Among them, the The fracture energy within the subdomain; This represents the length of the grid edge within the subdomain; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. The fracture energy at which the PyC subdomain fails to overcome van der Waals forces; The direction of the mesh edge and the weak surface The included angle.

[0055] Optionally, step 23 includes:

[0056] The fracture energy of each inter-subdomain grid edge is calculated using the following formula:

[0057] ;

[0058] Among them, the The fracture energy between subdomains; This represents the length of the grid edge between the subdomains; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. It is the smaller value of the angle between the grid edge and the weak direction of the two adjacent subdomains.

[0059] This application also provides a graph-based device for modeling and analyzing the debonding behavior of pyrolytic carbon interface phases, the graph-based device for modeling and analyzing the debonding behavior of pyrolytic carbon interface phases includes:

[0060] A model acquisition module is used to acquire a PyC interface phase model with SiC constraints.

[0061] The fracture energy assignment module is used to assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy.

[0062] A global crack propagation path acquisition module is used to acquire the global crack propagation path within the PyC interface phase based on a PyC mesh model in which all mesh edges are assigned corresponding fracture energies.

[0063] A key interface parameter calculation module is used to calculate key interface parameters based on the global crack propagation path within the PyC interface phase.

[0064] This application has the following advantages:

[0065] (1) This invention can establish and efficiently solve PyC graphical models with a large number of different oriented subdomains, and is particularly suitable for materials such as PyC with significant layered subdomain structures and strong anisotropic fracture behavior. Crack propagation within the PyC interface phase is mainly controlled by the orientation of the layered subdomains and the strong anisotropic fracture energy, while the influence of anisotropic constitutive model on the crack path is relatively minor. Based on this physical characteristic, this invention adopts a graph theory-based simulation method to transform the crack propagation problem into a minimum cut / energy optimization problem in the graphical model, avoiding explicit solutions to complex anisotropic constitutive models and nonlinear fracture evolution equations. Thus, this invention can simulate crack propagation in PyC interface phases containing tens of thousands or even more subdomains and obtain the crack path with global energy minimization. In contrast, traditional numerical methods for continuous media such as the finite element method and phase field method usually require high spatial discretization resolution and a large number of nonlinear solution iterations when dealing with structures containing a large number of anisotropic subdomains. The computational cost is high, the convergence difficulty is large, and it is difficult to efficiently perform large-scale, multi-subdomain interface phase crack path calculations.

[0066] (2) This invention proposes a moving window method suitable for simulating crack propagation in long and narrow interface phases. This method can dynamically update parameters such as the local computational region, crack tip position, and stress intensity factor during the gradual crack propagation process, thereby overcoming the problem that traditional graph theory simulation methods are usually only applicable to fixed computational regions and are difficult to describe the propagation of long-distance continuous cracks. Through the moving window method, this invention retains the advantages of global energy optimization of graph theory methods, avoids the high computational cost of solving the entire long-scale interface phase region at once, and considers the changes in crack tip and stress intensity factor caused by the continuous crack propagation, making it possible to simulate the debonding behavior of PyC interface phases with lengths of a few micrometers or even tens of micrometers.

[0067] (3) This invention can not only obtain the crack propagation path within the PyC interface phase, but also further extract key interface parameters such as interface fracture energy and interface roughness based on the global geometric features and energy information of the crack path. Compared with simulation methods that can only provide local crack deflection, this invention can establish the connection between "interface phase microstructure and interface performance parameters" by simulating the global propagation behavior of cracks within the interface phase on a larger scale, providing an effective means for characterizing the debonding behavior of the interface phase, optimizing the interface design of ceramic matrix composites, and obtaining parameters for multi-scale damage models. Attached Figure Description

[0068] Figure 1 This is a simplified flowchart illustrating a graph-theory-based method for simulating the debonding behavior of PyC interface phases in one embodiment of the present invention.

[0069] Figure 2 This is an example schematic diagram of a PyC discrete mesh model containing SiC boundaries.

[0070] Figure 3 This is a schematic diagram of the computational framework of the moving window method. Sub-figure ac illustrates the computational process of the moving window method, sub-figure d shows the setting of boundary conditions within a window, and sub-figure e shows the calculation of the stress intensity factor for the inclined crack.

[0071] Figure 4 This is an example schematic diagram showing the simulation results of crack propagation paths within the PyC interface phase with different OA values ​​and thicknesses.

[0072] Figure 5 This is a schematic diagram showing the changes in interfacial fracture energy and interfacial roughness with the thickness of the interfacial phase under different OA values. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0074] like Figure 1 The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces includes:

[0075] Step 1: Obtain the PyC interface phase model with SiC constraints;

[0076] Step 2: Assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy.

[0077] Step 3: Obtain the global crack propagation path within the PyC interface phase based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy;

[0078] In this embodiment, step 3 includes:

[0079] Step 31: Calculate the crack propagation path within the first window based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy;

[0080] Step 32: Obtain the global crack propagation path within the PyC interface phase based on the crack propagation path within the first window and the PyC mesh model where all mesh edges are assigned corresponding fracture energies.

[0081] Step 4: Calculate the key interface parameters based on the global crack propagation path within the PyC interface phase.

[0082] In this embodiment, step 1 includes:

[0083] Step 11: Obtain the complete set of basic feature parameters for the PyC interface phase to be modeled;

[0084] In this embodiment, the complete set of basic feature parameters for the PyC interface phase to be modeled includes:

[0085] The length and thickness of the PyC interface phase, and the two sides ( Figure 2 The thickness of the SiC layer (shown on both sides), the number of subdomains inside PyC, the orientation angle OA value of PyC, the shear modulus and Poisson's ratio of PyC, and the fracture energy of PyC subdomains overcoming chemical bond failure and overcoming van der Waals force failure.

[0086] Step 12: Define the geometric region of the PyC interface phase based on the complete set of basic feature parameters of the PyC interface phase to be modeled, thereby obtaining a two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries; for example, define the PyC interface phase as a region of length L. x = 1000 nm, thickness L y A two-dimensional rectangular region of 150 nm.

[0087] Step 13: Generate a set of N parameters based on the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries and the complete set of basic feature parameters of the PyC interface phase to be modeled. s PyC subdomain topology of a Voronoi polygon;

[0088] Specifically, based on the number of subdomains within PyC, a corresponding number of seed points are randomly seeded within the PyC region. Based on the locations of these seed points, Voronoi structures are generated within the PyC region, with each Voronoi polygon representing a subdomain of PyC. During modeling, the seed points are distributed as evenly as possible to avoid subdomains that are too large or too small.

[0089] Step 14: Generate an acute-angled triangular mesh set covering the entire PyC computational domain based on the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries and the complete set of basic feature parameters of the PyC interface phase to be modeled.

[0090] Specifically, based on the given number of grids, Delaunay triangulation is performed within the PyC region to generate a corresponding number of triangular grids. The number of grids can be roughly set based on the number of subdomains within PyC; empirically, the number of grids can be set to approximately 150 times the number of subdomains. This ensures simulation accuracy while avoiding excessive computational cost. Regarding grid shape, the goal is to ensure that the vast majority of grids are acute-angled triangles; any individual grids with sharp angles can be removed or optimized.

[0091] Step 15: Identify the subdomain to which each mesh belongs based on the PyC subdomain topology and the acute-angled triangle mesh set, thereby obtaining a set of triangle meshes with subdomain number labels;

[0092] Specifically, it is determined which subdomain the center of each grid cell is located within. Each grid cell is then numbered according to its subdomain number, and grid cells with the same number form a subdomain with the corresponding number.

[0093] Step 16: Perform smooth subdomain boundary optimization based on the triangular mesh set with subdomain number labels to obtain a PyC mesh model with smooth continuous subdomain boundaries;

[0094] Specifically, the method established by Siddhartha Srivastava et al. in their paper "Graph Coloring Approach to Mesh Generation in Multiphase Media with Smooth Boundaries" can be used to update the numbering of mesh cells near the boundaries of each subdomain, so that subdomains composed of meshes with the same number have smoother boundaries.

[0095] Step 17: Generate a PyC mesh model with smooth, continuous subdomain boundaries and PyC OA values, where each subdomain has a unique weak surface direction. The PyC mesh model, wherein each subdomain has a unique weak surface direction. The PyC mesh model is used as the discrete mesh model of the PyC interface phase;

[0096] Specifically, the orientation of each subdomain is determined based on the OA value of PyC. Perform random assignment to ensure that the value follows a Gaussian distribution:

[0097] Formula 1;

[0098] in, Let be the probability density function of the orientation angle of the subdomain; The orientation angle of the weak surface direction of the subdomain; Let be the standard deviation of a Gaussian distribution, satisfying = OA / 2 (0≤OA≤π); OA is the orientation angle of PyC; obviously, the larger OA is, the greater the orientation dispersion of the subdomain.

[0099] Step 18: Discretize the mesh model on both sides of the PyC interface ( Figure 2 The model introduces SiC constraints into the upper and lower layers (as shown), thereby obtaining a PyC interface phase model with SiC constraints. Specifically, the SiC above and below PyC is introduced into the model as the boundary for crack propagation. The fracture energy corresponding to the mesh edge at the SiC and PyC interface is set to a maximum value (e.g., 10000 J / m). 2 This ensures that cracks can only propagate within the PyC region. At the boundary between the SiC and PyC regions, a shared-node, shared-edge mesh connection method is used, meaning that each mesh edge on the boundary line corresponds one-to-one and fits continuously, without mesh misalignment, overlap, or gaps.

[0100] Specifically, SiC regions are generated above and below PyC. The SiC regions above and below PyC are introduced into the model as boundaries for crack propagation. If crack propagation within SiC is not considered, the mesh within SiC can be relatively sparse. It is necessary to ensure that the mesh edges at the interface between PyC and SiC are strictly aligned.

[0101] Figure 2 An example of a PyC discrete mesh model with SiC boundary constraints is presented. In this model, the PyC length is 1000 nm and the thickness is 150 nm, while the SiC thickness is 10 nm. The number of PyC subdomains is approximately 13,700, and the OA value is 100°. The total number of meshes in the model is approximately 2,040,000.

[0102] In this embodiment, step 2 includes:

[0103] Step 21: Extract all mesh edges of the PyC interface phase model with SiC constraints and classify them to obtain the set of classified mesh edges;

[0104] Step 22: Calculate the fracture energy of the mesh edges in the subdomain based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all mesh edges in the subdomain;

[0105] In this embodiment, step 22 includes:

[0106] The fracture energy of each mesh edge within a subdomain (i.e., mesh edges where mesh cells on both sides belong to the same subdomain) is calculated using the following formula:

[0107] ;

[0108] Among them, the The fracture energy within the subdomain; This represents the length of the grid edge within the subdomain; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. The fracture energy at which the PyC subdomain fails to overcome van der Waals forces; Let α be the angle between the edge of the mesh and the direction of the weak surface of the subdomain, where α ∈ [0, π / 2]. Based on experimental results, and 14 J / m 2 and 0.39 J / m 2 .

[0109] Step 23: Calculate the fracture energy of the inter-subdomain mesh edges based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all inter-subdomain mesh edges;

[0110] In this embodiment, step 23 includes:

[0111] The fracture energy of each inter-subdomain grid edge is calculated using the following formula:

[0112] ;

[0113] in, The fracture energy at the grid edge between subdomains; This represents the length of the grid edge between the subdomains; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. The smaller value is the angle between the grid edge and the weak direction of the two adjacent subdomains. ∈ [0, π / 2].

[0114] The fracture energy values ​​corresponding to the mesh edges within all subdomains and the fracture energy values ​​corresponding to the mesh edges between all subdomains are assigned fracture energy attributes to the classified mesh edge set, thereby obtaining a PyC interface phase model with SiC constraints in which all mesh edges are assigned corresponding fracture energy attributes.

[0115] In this embodiment, step 3 includes:

[0116] Extract the first computational window of the PyC mesh model where all mesh edges are assigned the corresponding fracture energy, thereby obtaining the independent computational domain of the first window;

[0117] In this embodiment, the leftmost extraction width of the PyC mesh model, where all mesh edges are assigned the corresponding fracture energy, is... The region is used to obtain the independent computational domain of the first window;

[0118] Initial crack boundary conditions and external load conditions are set for the independent computational domain of the first window, thereby obtaining the computational domain of the first window with fixed initial crack boundary conditions.

[0119] The crack tip stress intensity factor is calculated based on the first window computational domain with fixed initial crack boundary conditions and external load conditions, thereby obtaining the crack tip type I stress intensity factor and crack tip type II stress intensity factor.

[0120] The energy release rate as a function of the crack deflection angle is obtained based on the type I and type II stress intensity factors at the crack tip.

[0121] The set of cost function values ​​for all mesh elements within the first window is generated based on the energy release rate as a function of the crack deflection angle and the first window computational domain with fixed initial crack boundary conditions.

[0122] Obtain the total cost function based on the set of cost function values ​​of all grid cells in the first window;

[0123] Based on the total cost function and the first window computational domain with fixed initial crack boundary conditions, constrained graph cut optimization is performed to find the minimum cost crack path, thereby obtaining the initial crack propagation path within the first window.

[0124] Quasi-static expansion control is applied to the initial crack propagation path within the first window to obtain the complete crack propagation path within the first window that satisfies the quasi-static expansion condition.

[0125] In this embodiment, step 32 includes:

[0126] Step 321: From the complete crack propagation path within the first window that satisfies the quasi-static propagation condition, extract a preset length after the initial crack tip. The crack segment is added to the global crack propagation path; a preset length is then cut off from the tail end of the crack segment along the crack propagation direction. The crack segment is used as the initial crack in the second calculation window, and the inclination angle of the initial crack in the second window is calculated. In this embodiment, from the complete crack propagation path within the first window that satisfies the quasi-static propagation condition, a preset length is taken after the right endpoint of the initial crack. The crack segment is added to the global crack propagation path; a preset length is then cut off from the tail end of the crack segment along the crack propagation direction. The crack segment is used as the initial crack in the second calculation window, and the inclination angle of the initial crack in the second window is calculated.

[0127] Step 322: Translate the first calculation window along the direction of macroscopic crack propagation until the window boundary coincides with the initial crack initiation point of the second calculation window, thus creating the independent calculation domain of the second calculation window;

[0128] In this embodiment, the starting point of the initial crack in the second calculation window is used as the left boundary. (It can be understood that the position of the second window is related to the macroscopic direction of the crack. When the macroscopic direction of the crack is from left to right, the second window is located to the right of the first window; conversely, if it is from right to left, the second window is located to the left of the first window.) The extracted width is... The model region is used as an independent computational domain for the second computational window;

[0129] Step 323: Within the independent computational domain of the second computational window, solve for the complete crack propagation path that satisfies the quasi-static propagation condition within the second window using the same computational method as the first window;

[0130] Step 324: Repeat the following operation until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. In this embodiment, the following operations are performed until the distance between the left boundary of the window to be generated and the right boundary of the PyC interface phase model containing SiC constraints is less than or equal to a preset threshold. :

[0131] Step 325: From the complete crack propagation path of the currently solved k-th window, extract the length after the crack tip of the initial crack in the k-th window, which is... The crack segment is added to the global crack propagation path; where, In this embodiment, from the complete crack propagation path of the currently solved k-th window, the length after the right endpoint of the initial crack is extracted. The crack segment is added to the global crack propagation path; where, ;

[0132] Step 326: Cut a length of [length missing] from the tail end of the crack segment in the opposite direction. The crack segment is used as the initial crack in the (k+1)th calculation window, and its dip angle is calculated.

[0133] Step 327: Following the methods in steps 322 to 323, generate the independent computational domain of the (k+1)th computational window and solve for the complete crack propagation path that satisfies the quasi-static propagation condition;

[0134] Step 328: Continue until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. In this embodiment, "time" refers to the time when the distance between the left boundary of the window and the right boundary of the PyC interface phase model containing SiC constraints is less than or equal to a preset threshold. (At that time), generate the last calculation window, and solve for the complete crack propagation path of the last window according to the methods in steps 322 to 323;

[0135] Wherein, if the length obtained during the extraction of the crack segment in any window is... If the absolute value of the initial crack inclination angle is greater than or equal to 90°, then the crack extension segment entering the global crack propagation path is lengthened, and the tail length is re-trunculated. The crack segment is used as the initial crack for the next window until the absolute value of the initial crack dip angle is less than 90°.

[0136] Step 329: Piece together all the crack segments in the global crack propagation path according to the window progression order to obtain the global crack propagation path within the PyC interface.

[0137] In general, the idea of ​​calculating crack propagation within a narrow interface phase using the moving window method in this application is as follows: Figure 3 As shown, the specific steps are as follows:

[0138] First, a region with a width of L0 and a thickness equal to the sum of the thicknesses of PyC and SiC on both sides is delineated on the left side of the PyC interface phase model containing SiC constraints as the first window.

[0139] In the middle left of the first window, set a length of... Given a horizontal initial crack, the complete crack path within the first window is calculated based on the initial crack and external load using the principle of energy minimization.

[0140] In this complete crack path, the length after retaining the initial crack is... The crack segment enters the global crack path; within this length... In the extended segment of the crack, the length of its tail is taken as... The crack segment is retained and enters the next window, serving as the initial crack for the next window.

[0141] Next, the window is shifted to the right so that its left boundary coincides with the left endpoint of the initial crack segment inherited from the previous window, resulting in the second window. Similarly, the complete crack path within the second window and the crack extension segments retained in the global crack path are calculated. And the length of the end of the extension segment retained in the third window as the initial crack is... The crack segment. The window is pushed to the right sequentially like this. When the distance between the left boundary of the window and the right boundary of the model is less than L1, it is considered to have entered the last window, and the complete crack path within that window is directly retained and entered into the global crack path.

[0142] Thus, by using the moving window method, the change in the crack tip stress intensity factor caused by the continuous propagation of the crack within the narrow interface phase can be effectively considered, thereby calculating the global crack propagation path more accurately. In this embodiment, L0 is set to 150 nm. The wavelength is set to 10nm. The wavelength is set to 50nm, and L1 is set to 70nm.

[0143] These parameter values ​​can be adjusted according to the model size, selecting a suitable value between computational accuracy and computational efficiency.

[0144] It is worth noting that a relatively rare situation requiring special handling may arise during implementation: a 50nm long crack segment within a certain window that is retained as part of the global crack path may have a 10nm long crack segment at its tail end with an inclination angle greater than or equal to 90°, making it impossible to retain it as the initial crack in the next window. In this case, the crack extension segment within the window that is retained as part of the global crack path is lengthened, for example, to 60nm, and then the 10nm long portion at the tail end of this 60nm segment is retained as the initial crack in the next window.

[0145] In this embodiment, the crack path is calculated based on the principle of global energy minimization:

[0146] Equation 4,

[0147] in, Let G be the total energy, Γ be the crack path, and G be the crack path. c Where G is the fracture energy and G is the energy release rate. Let be the infinitesimal length of the crack path. In graph theory-based methods, energy is equivalent to the cost of labeling mesh elements. Finding the graph cut path that minimizes energy is essentially finding the way to label mesh elements with the minimum cost. The mesh edges between all adjacent mesh elements with different labels form the crack path.

[0148] The total cost of labeling a grid cell is:

[0149] Equation 5,

[0150] in, For the total cost, For the first One grid cell, For the first One grid cell, and They are grid cells The centroid and label (1 or 2). ( , ) is a function related to the energy release rate. It is a unit The area; For unit and unit The length of the grid edges between them This represents the fracture energy corresponding to the edge of the grid. δ x This is a Dirac-delta function, equal to 1 at x = 0 and 0 elsewhere. This means that only when and The fracture energy corresponding to the mesh edge between them is calculated only when the labels are different, that is, when the crack passes through this mesh edge.

[0151] For the PyC interface phase, strong fracture energy anisotropy within the subdomain is the dominant factor controlling crack deflection, while constitutive anisotropy is only a secondary factor. Therefore, in the graph theory-based simulation method, the constitutive anisotropy of PyC is simplified, and an isotropic constitutive model is used for calculation. For a crack of length 2a and dip angle β in an isotropic material ( Figure 3 e), if subjected to a pure shear load, i.e., σ xx = σ yy = 0 and σ xy = τ, then the crack tip stress intensity factor is:

[0152] ;

[0153] Formula 6.

[0154] in, Type I stress intensity factor at the crack tip; Type II stress intensity factor at the crack tip; Shear stress; The angle of inclination of the crack; The initial crack length within the window; Pi is the mathematical constant of a circle.

[0155] In the moving window method, the crack tip stress intensity factor within each window is calculated according to Equation 6. Additionally, if If < 0, then pass the command = 0 to ignore Contribution to energy release rate.

[0156] In each window, the initial crack is formed by fixing the mesh cell label above it to 1 and the mesh cell label below it to 2, as shown below. Figure 3 As shown in d. Ignoring anisotropic constitutive properties, the energy release rate of the crack deflection angle ω relative to the initial crack direction is:

[0157] Equation 7,

[0158] in, The energy release rate is the angle ω of the crack deflection relative to the initial crack direction. θ is the deflection angle of the crack relative to the initial crack direction; μ is the shear modulus. for Type stress intensity factor, for Type stress intensity factor, For parameters related to Poisson's ratio, under plane strain conditions κ = 3−4ν, where ν is Poisson's ratio. For elements within the window... ,function It can be determined by the following formula:

[0159] Formula 8,

[0160] in, This is a function related to the energy release rate at the centroid of the grid cells. and They are grid cells The centroid and label (1 or 2). The angular distribution function related to the energy release rate. and It is the center of mass. Polar angle and polar radius in the initial crack coordinate system The following relationship exists between the energy release rate G and the energy release rate G:

[0161] Formula 9,

[0162] From Equation 9, we can obtain:

[0163] ;

[0164] Formula 10

[0165] Where, the expression for c is:

[0166] Formula 11.

[0167] in, The angular distribution function corresponding to label 1; The angular distribution function corresponding to label 2; Polar angle; The initial crack inclination angle; The crack deflection angle is Energy release rate at that time; The angle of deflection of the crack relative to the initial crack direction.

[0168] The angular distribution function corresponding to label 1 at the deflection angle The value at; The angular distribution function corresponding to label 2 at the deflection angle The value at; Energy release rate For deflection angle The derivative; Here, c is the Heaviside step function; c is a constant term.

[0169] It is a function of energy release rate; The initial crack inclination angle; Energy release rate versus angle The derivative; The integral variable is the angle.

[0170] In summary, the total cost function in Equation 5 can be calculated based on Equations 2-3 and 6-11. Using this, the graph cut optimization library (gco library) developed by Veksler and Delong et al. can efficiently calculate the mesh element labeling method with the minimum cost, thereby obtaining the corresponding crack path.

[0171] In this embodiment, step 4 includes:

[0172] Step 41: Based on the global crack path, sum the fracture energy corresponding to each crack micro-segment in the path, and then divide by the apparent length of the interface phase (1000 nm in this embodiment) to obtain the interface fracture energy.

[0173] Step 42: Calculate the interface roughness based on the global crack propagation path within the PyC interface phase, thereby obtaining the interface roughness of the PyC interface phase.

[0174] In this embodiment, the interface roughness can be calculated based on the coordinates of the endpoints of all crack micro-segments. :

[0175] Equation 12,

[0176] Where N is the total number of endpoints. This is the average of the coordinates of all endpoints in the y-direction (thickness direction). Let y be the y-coordinate of the i-th endpoint. Figure 5 The variations of interfacial fracture energy and interfacial roughness with interfacial phase thickness under different OA values ​​are shown. It can be seen that the interfacial fracture energy first decreases and then tends to remain constant with increasing thickness. For high-texture PyC, the interfacial roughness changes little with thickness; for medium and low-texture PyC, the interfacial roughness first increases and then tends to remain constant with thickness. The simulation results validate the experimental results.

[0177] In this embodiment, the total energy release rate generated by the combination of external shear stress τ, interfacial shear modulus μ, and Poisson's ratio ν is slightly greater than the total fracture energy to ensure that the crack propagation is approximately quasi-static. Specifically, the total energy (cost) in control equation 5 is less than 0 but greater than an empirical value (taken as –200 × 10⁻⁶). –9 The crack propagation is guaranteed by J / m. In this case, crack propagation is dominated by strongly anisotropic fracture energy, so the individual values ​​of τ, μ, and ν are not of practical reference value and can be reasonably adjusted in the process of controlling the total energy. Figure 4 The simulations show crack propagation paths within the interface phases of different OA values ​​and thicknesses. In the simulations, the Poisson's ratio was set to 0.25 for all models, and the shear modulus was set to 500, 1000, 1600, and 1900 MPa for OA = 10°, 40°, 70°, and 100°, respectively. The total energy was controlled to be less than 0 but greater than -200 × 10⁻⁶ MPa by adjusting the external load τ. –9 J / m is used to ensure quasi-static expansion.

[0178] This invention provides a graph-theory-based method for modeling PyC interface phases and analyzing their debonding behavior. First, a discrete mesh model of the interface phase is established based on the given interface phase size, number of subdomains, and OA value. Then, based on atomic-scale simulations or experimental results, fracture energy is assigned to each mesh edge in the PyC mesh model. Next, using the moving window method, crack propagation paths are obtained within each window based on the energy minimization principle. By advancing the window, the global crack path is calculated. Finally, based on the global crack path, key interface parameters such as interface fracture energy and interface roughness are calculated. This invention can provide valuable guidance for the study of PyC fracture behavior and the interface design of ceramic matrix composites.

[0179] This application also provides a graph-based device for modeling and analyzing the debonding behavior of pyrolytic carbon interface phases, the graph-based device for modeling and analyzing the debonding behavior of pyrolytic carbon interface phases includes:

[0180] A model acquisition module is used to acquire a PyC interface phase model with SiC constraints.

[0181] The fracture energy assignment module is used to assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy.

[0182] A global crack propagation path acquisition module is used to acquire the global crack propagation path within the PyC interface phase based on a PyC mesh model in which all mesh edges are assigned corresponding fracture energies.

[0183] A key interface parameter calculation module is used to calculate key interface parameters based on the global crack propagation path within the PyC interface phase.

[0184] This application has the following advantages:

[0185] (1) This invention can establish and efficiently solve PyC graphical models with a large number of different oriented subdomains, and is particularly suitable for materials such as PyC with significant layered subdomain structures and strong anisotropic fracture behavior. Crack propagation within the PyC interface phase is mainly controlled by the orientation of the layered subdomains and the strong anisotropic fracture energy, while the influence of anisotropic constitutive model on the crack path is relatively minor. Based on this physical characteristic, this invention adopts a graph theory-based simulation method to transform the crack propagation problem into a minimum cut / energy optimization problem in the graphical model, avoiding explicit solutions to complex anisotropic constitutive models and nonlinear fracture evolution equations. Thus, this invention can simulate crack propagation in PyC interface phases containing tens of thousands or even more subdomains and obtain the crack path with global energy minimization. In contrast, traditional numerical methods for continuous media such as the finite element method and phase field method usually require high spatial discretization resolution and a large number of nonlinear solution iterations when dealing with structures containing a large number of anisotropic subdomains. The computational cost is high, the convergence difficulty is large, and it is difficult to efficiently perform large-scale, multi-subdomain interface phase crack path calculations.

[0186] (2) This invention proposes a moving window method suitable for simulating crack propagation in long and narrow interface phases. This method can dynamically update parameters such as the local computational region, crack tip position, and stress intensity factor during the gradual crack propagation process, thereby overcoming the problem that traditional graph theory simulation methods are usually only applicable to fixed computational regions and are difficult to describe the propagation of long-distance continuous cracks. Through the moving window method, this invention retains the advantages of global energy optimization of graph theory methods, avoids the high computational cost of solving the entire long-scale interface phase region at once, and considers the changes in crack tip and stress intensity factor caused by the continuous crack propagation, making it possible to simulate the debonding behavior of PyC interface phases with lengths of a few micrometers or even tens of micrometers.

[0187] (3) This invention can not only obtain the crack propagation path within the PyC interface phase, but also further extract key interface parameters such as interface fracture energy and interface roughness based on the global geometric features and energy information of the crack path. Compared with simulation methods that can only provide local crack deflection, this invention can establish the connection between "interface phase microstructure and interface performance parameters" by simulating the global propagation behavior of cracks within the interface phase on a larger scale, providing an effective means for characterizing the debonding behavior of the interface phase, optimizing the interface design of ceramic matrix composites, and obtaining parameters for multi-scale damage models.

[0188] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A graph-based pyrocarbon interfacial phase modeling and debonding behavior analysis method, characterized in that, The graph theory-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces includes: Step 1: Obtain the PyC interface phase model with SiC constraints; Step 2: Assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy. Step 3: Obtain the global crack propagation path within the PyC interface phase based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy; Step 4: Calculate the key interface parameters based on the global crack propagation path within the PyC interface phase.

2. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 1, characterized in that, Step 3 includes: Step 31: Calculate the crack propagation path within the first window based on the PyC mesh model where all mesh edges are assigned the corresponding fracture energy; Step 32: Obtain the global crack propagation path within the PyC interface phase based on the crack propagation path within the first window and the PyC mesh model where all mesh edges are assigned corresponding fracture energies.

3. The graph theory-based method for modeling pyrolytic carbon interface phases and analyzing debonding behavior as described in claim 2, characterized in that, Step 1 includes: Step 11: Obtain the complete set of basic feature parameters for the PyC interface phase to be modeled; Step 12: Define the geometric region of the PyC interface phase based on the complete set of basic feature parameters of the PyC interface phase to be modeled, thereby obtaining the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries. Step 13: Generating PyC sub-domain topology containing N s Voronoi polygons from the PyC interface phase two-dimensional rectangular computational domain with coordinate boundaries and the complete set of base characteristic parameters of the PyC interface phase to be modeled; Step 14: Generate an acute-angled triangular mesh set covering the entire PyC computational domain based on the two-dimensional rectangular computational domain of the PyC interface phase with coordinate boundaries and the complete set of basic feature parameters of the PyC interface phase to be modeled. Step 15: Identify the subdomain to which each mesh belongs based on the PyC subdomain topology and the acute-angled triangle mesh set, thereby obtaining a set of triangle meshes with subdomain number labels; Step 16: Perform smooth subdomain boundary optimization based on the triangular mesh set with subdomain number labels to obtain a PyC mesh model with smooth continuous subdomain boundaries; Step 17: Generate a unique weak surface direction for each subdomain based on the complete set of fundamental feature parameters of the PyC mesh model with smooth continuous subdomain boundaries and the PyC interface phase to be modeled. The PyC mesh model, wherein each subdomain has a unique weak surface direction. The PyC mesh model is used as the discrete mesh model of the PyC interface phase; Step 18: Introduce SiC constraints on both sides of the PyC interface phase discrete mesh model to obtain a PyC interface phase model with SiC constraints.

4. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 3, characterized in that, Step 2 includes: Step 21: Extract all mesh edges of the PyC interface phase model with SiC constraints and classify them to obtain the set of classified mesh edges; Step 22: Calculate the fracture energy of the mesh edges in the subdomain based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all mesh edges in the subdomain; Step 23: Calculate the fracture energy of the inter-subdomain mesh edges based on the classified mesh edge set, thereby obtaining the fracture energy values ​​corresponding to all inter-subdomain mesh edges; The fracture energy values ​​corresponding to the mesh edges within all subdomains and the fracture energy values ​​corresponding to the mesh edges between all subdomains are assigned fracture energy attributes to the classified mesh edge set, thereby obtaining a PyC interface phase model with SiC constraints in which all mesh edges are assigned corresponding fracture energy attributes.

5. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 4, characterized in that... Step 31 includes: Extract the first computational window of the PyC mesh model where all mesh edges are assigned the corresponding fracture energy, thereby obtaining the independent computational domain of the first window; Initial crack boundary conditions and external load conditions are set for the independent computational domain of the first window, thereby obtaining the computational domain of the first window with fixed initial crack boundary conditions. The crack tip stress intensity factor is calculated based on the first window computational domain with fixed initial crack boundary conditions and external load conditions, thereby obtaining the crack tip type I stress intensity factor and crack tip type II stress intensity factor. The energy release rate as a function of the crack deflection angle is obtained based on the type I and type II stress intensity factors at the crack tip. The set of cost function values ​​for all mesh elements within the first window is generated based on the energy release rate as a function of the crack deflection angle and the first window computational domain with fixed initial crack boundary conditions. Obtain the total cost function based on the set of cost function values ​​of all grid cells in the first window; Based on the total cost function and the first window computational domain with fixed initial crack boundary conditions, a constrained graph cut optimization is performed to find the minimum cost crack path, thereby obtaining the initial crack propagation path within the first window. Quasi-static propagation control is applied to the initial crack propagation path within the first window to obtain the complete crack propagation path within the first window that satisfies the quasi-static propagation conditions.

6. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 5, characterized in that, Step 32 includes: Step 321: From the complete crack propagation path within the first window that satisfies the quasi-static propagation condition, extract a preset length after the initial crack tip. The crack segment is added to the global crack propagation path; a preset length is then cut off from the tail end of the crack segment along the crack propagation direction. The crack segment is used as the initial crack in the second calculation window, and the inclination angle of the initial crack in the second window is calculated. Step 322: Translate the first calculation window along the direction of macroscopic crack propagation until the window boundary coincides with the initial crack initiation point of the second calculation window, thus creating the independent calculation domain of the second calculation window; Step 323: Within the independent computational domain of the second computational window, solve for the complete crack propagation path that satisfies the quasi-static propagation condition within the second window using the same computational method as the first window; Step 324: Repeat the following operation until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. : Step 325: From the complete crack propagation path of the currently solved k-th window, extract the length after the endpoint of the initial crack in the k-th window, which is... The crack segment is added to the global crack propagation path; where, ; Step 326: Cut a length of [length missing] from the tail end of the crack segment in the opposite direction. The crack segment is used as the initial crack in the (k+1)th calculation window, and its dip angle is calculated. Step 327: Following the methods in steps 322 to 323, generate the independent computation domain of the (k+1)th computation window and solve for the complete crack propagation path that satisfies the quasi-static propagation condition; Step 328: Continue until the distance between the initial crack initiation point of the current window and the model's terminal boundary is less than or equal to a preset threshold. When the last calculation window is obtained, the complete crack propagation path of the last window is solved according to the methods in steps 322 to 323; Wherein, if the length obtained during the extraction of the crack segment in any window is... If the absolute value of the initial crack inclination angle is greater than or equal to 90°, then the crack extension segment entering the global crack propagation path is lengthened, and the tail length is re-trunculated. The crack segment is used as the initial crack for the next window until the absolute value of the initial crack dip angle is less than 90°. Step 329: Piece together all the crack segments in the global crack propagation path according to the window progression order to obtain the global crack propagation path within the PyC interface.

7. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 6, characterized in that, Step 4 includes: Step 41: Calculate the interfacial fracture energy based on the global crack propagation path within the PyC interface phase, thereby obtaining the interfacial fracture energy of the PyC interface phase. Step 42: Calculate the interface roughness based on the global crack propagation path within the PyC interface phase, thereby obtaining the interface roughness of the PyC interface phase.

8. The method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces based on graph theory as described in claim 7, characterized in that, Step 22 includes: The fracture energy of each mesh edge within a subdomain is calculated using the following formula: ; Among them, the The fracture energy within the subdomain; This represents the length of the grid edge within the subdomain; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. The fracture energy at which the PyC subdomain fails to overcome van der Waals forces; The orientation angle of the grid edge and the orientation of the weak surface The included angle.

9. The graph-based method for modeling and analyzing the debonding behavior of pyrolytic carbon interfaces as described in claim 8, characterized in that, Step 23 includes: The fracture energy of each inter-subdomain grid edge is calculated using the following formula: ; in, The fracture energy at the grid edge between subdomains; This represents the length of the grid edge between the subdomains; This represents the breaking energy at which the PyC subdomains fail to overcome chemical bond interactions. It is the smaller value of the angle between the grid edge and the weak direction of the two adjacent subdomains.

10. A graph-theory-based apparatus for modeling pyrolytic carbon interface phases and analyzing debonding behavior, characterized in that, The graph theory-based pyrolysis carbon interface phase modeling and debonding behavior analysis device includes: A model acquisition module is used to acquire a PyC interface phase model with SiC constraints. The fracture energy assignment module is used to assign fracture energy to all mesh edges of the obtained PyC interface phase model with SiC constraints, thereby obtaining a PyC mesh model in which all mesh edges are assigned the corresponding fracture energy. A global crack propagation path acquisition module is used to acquire the global crack propagation path within the PyC interface phase based on a PyC mesh model in which all mesh edges are assigned corresponding fracture energies. A key interface parameter calculation module is used to calculate key interface parameters based on the global crack propagation path within the PyC interface phase.