A damage prediction method for three-dimensional woven ceramic matrix composites containing randomly varying pore defects

Abstract

The present invention discloses a method for predicting damage of a three-dimensional braided ceramic matrix composite containing randomly varying pore defects. Whether in engineering applications or scientific research, three-dimensional braided ceramic matrix composites are at the forefront of research on aerospace high-temperature thermal structure materials, and the mechanical properties of three-dimensional braided ceramic matrix composites depend on their mesoscopic structure and porosity content. The method of the present invention takes into account the state characteristics of pore defects generated during the forming process of ceramic matrix composites, extracts the mesoscopic morphology state information of the material, conducts clustering analysis based on pore area, diameter size, and position, uses finite element analysis to perform mesoscopic unit cell modeling of the material, and embeds the damage analysis theory into the finite element model through the method of user-defined subroutines for iterative calculation, thereby realizing efficient prediction calculation of the damage of three-dimensional braided ceramic matrix composites.

Description

Technical Field

[0001] The present invention belongs to the technical field of data prediction and simulation analysis of woven composites, and relates to a damage prediction method for three-dimensional woven ceramic matrix composites containing randomly varying pore defects. Background Art

[0002] The three-dimensional woven structure does not require stitching and machining, and can be directly integrally woven into preform components of any complex shape, with strong designability. In addition, the space-interlocked network structure of the three-dimensional woven fabric also fundamentally makes up for some mechanical property deficiencies of traditional laminated composites, such as low interlaminar shear strength and low damage tolerance. The maximum working temperature of the hot-end components of a space engine based on three-dimensional woven ceramic matrix composites can be increased by about 300 - 500 °C compared with conventional composites, enabling the hot-end components to serve in a higher temperature environment to meet the "thermal" demand for the increase in the specific impulse of the engine; on the premise that the weight of the hot-end components of a space engine based on three-dimensional woven ceramic matrix composites is reduced by 50% - 70%, the resistance to damage and crack propagation performance can be effectively improved compared with conventional composites to meet the working performance "force" demand for the increase in the specific impulse of the engine. As a new type of high-temperature resistant structural composite material, three-dimensional woven ceramic matrix composites have broad application prospects in the fields of high thrust-to-weight ratio aeroengines, liquid and solid rocket engines, thermal protection systems of aerospace vehicles, nuclear reactors, etc.

[0003] Using a three-dimensional determinant weaving machine to weave a three-dimensional preform based on the four-step method is currently the most commonly used weaving method. In each weaving cycle, the yarn carriers on the chassis will move four steps, and the distance of each step is equal, which is called the four-step 1×1 weaving process. The precursor infiltration and pyrolysis method (PIP) uses a fiber preform as the skeleton, and the precursor is infiltrated into the preform by vacuum or pressure and then crosslinked and cured, and then pyrolyzed at high temperature to convert the precursor polymer into a ceramic matrix. After multiple impregnation-crosslinking-pyrolysis densification processes, ultra-high temperature ceramic matrix composites are obtained. The PIP process is used to introduce ceramic phases into the matrix of the composite material. The process is as follows: The preform is vacuum impregnated with a precursor solution of a certain concentration, and then impregnated. After completion, the sample is taken out of the impregnation solution and baked in an oven. After drying, the sample is heat-treated in a graphitization furnace, held at a certain temperature for a certain time under a protective atmosphere, and then cooled to room temperature with the furnace. Finally, the sample is densified to the required density of the composite material through multiple impregnation and pyrolysis processes. However, during the impregnation and pyrolysis process, problems such as a large number of pore defects and uneven pore distribution are inevitable, resulting in a complex internal component structure of the three-dimensional woven ceramic matrix composites and difficult extraction of pore characteristics after forming. Summary of the Invention

[0004] The object of the present invention is to overcome the above-mentioned disadvantages of the prior art, and to provide a method for predicting damage of a three-dimensional braided ceramic matrix composite material with randomly varying pore defects, so as to solve the problems in the prior art that it is difficult to obtain the mesoscopic defect parameters of the three-dimensional braided ceramic matrix composite material; the pore distribution of the three-dimensional braided ceramic matrix composite material is uneven, showing a gradient distribution with more in the middle and less on both sides, and it is difficult to conduct finite element analysis research on the actual defects of the material, as well as the overall calculation efficiency is very low, which is severely limited by the size and shape of the mesh elements.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for predicting damage of a three-dimensional braided ceramic matrix composite material with randomly varying pore defects, comprising the following steps:

[0007] Step 1, obtain the pore defects inside the composite material through cross-section scanning;

[0008] Step 2, obtain the pore parameters, defect distribution and porosity inside the composite material through cluster analysis;

[0009] Step 3, establish a finite element mesoscopic model, and input the pore parameters, defect distribution and porosity into the finite element mesoscopic model;

[0010] Step 4, randomly generate pore units for the finite element mesoscopic model through the Monte Carlo algorithm to obtain a finite element model with pores;

[0011] Step 5, for the finite element model with pores, assign material properties and apply periodic boundary conditions and loads to obtain a finite element model with pores and boundary conditions and loads;

[0012] Step 6, use the damage analysis method to predict the finite element model with pores and boundary conditions and loads, and predict the damage and yield characteristics of the finite element model under various stress conditions;

[0013] Step 7, obtain the failure condition of the composite material according to the calculation results of the damage and yield characteristics.

[0014] A further improvement of the present invention lies in:

[0015] Preferably, in step 1, the cross-section is scanned by X-ray.

[0016] Preferably, in step 1, the composite material is cut into tetrahedrons before scanning.

[0017] Preferably, the specific process of step 2 is:

[0018] 2.1 Randomly select n pore centroids as the basic centroids;

[0019] 2.2 Calculate the Euclidean distance between each remaining pore and the centroid of the nearest base, form an array of the Euclidean distances related to each pore centroid, and calculate the new centroid of each array.

[0020] 2.3 Calculate the Euclidean distance between all pores and the nearest new centroid to each of them, and divide the distances related to each new centroid into the array of the new centroid.

[0021] 2.4 Repeat steps 2.2 and 2.3 to obtain the pore parameters and defect distribution inside the composite material until the centroid no longer changes or reaches the maximum number of iterations, and take the centroid of the finally obtained array as the defect coordinates of the composite material.

[0022] Preferably, the calculation formula of the Euclidean distance is:

[0023]

[0024] where x1, y1, z1 are the rectangular spatial coordinates of the centroid, and x2, y2, z2 are the rectangular spatial coordinates of the pore respectively.

[0025] Preferably, in step 5, the material property is a three-dimensional four-directional ceramic matrix composite, the woven matrix is C / SiC, and the woven yarn is a T800-12K woven yarn.

[0026] Preferably, in step 3, when dividing the predetermined placement area of the pores, the fiber bundle and the matrix are divided separately.

[0027] Preferably, in step 6, the damage analysis method of the matrix is the Christensen damage criterion.

[0028] Preferably, in step 6, the damage analysis method of the fiber bundle is the Hashin criterion, and the strength parameter adopts the Chamis micromechanical strength model.

[0029] Preferably, in step 7, the failure conditions include the damage degrees in three directions of the composite material, fiber debonding and shear failure of the matrix.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention discloses a method for predicting damage of three-dimensional braided ceramic matrix composites with randomly varying pore defects. Whether in engineering applications or scientific research, three-dimensional braided ceramic matrix composites are at the forefront of research on high-temperature thermal structure materials for aerospace. The mechanical properties of three-dimensional braided ceramic matrix composites depend on their mesoscopic structure and porosity content. The method of the present invention considers the state characteristics of pore defects generated during the forming process of ceramic matrix composites, extracts the mesoscopic morphology state information of the material, conducts clustering analysis based on pore area, diameter size, and position, uses finite element analysis to perform mesoscopic unit cell modeling of the material, and embeds the damage analysis theory into the finite element model through the method of user-defined subroutines for iterative calculation, thereby realizing efficient prediction and calculation of the damage of three-dimensional braided ceramic matrix composites.

[0032] Furthermore, in the present invention, X-ray scan data is pre-filtered and the gray threshold is adjusted to separate pores from other components, and pores of different sizes can be identified. The mesoscopic pore defect parameters of three-dimensional braided ceramic matrix composites are obtained through analysis.

[0033] Furthermore, in the present invention, the internal pores of the material are subjected to clustering analysis to obtain the internal pore parameters of the material and the statistical situation of the defect distribution of the material, and the pore density regions of the fiber matrix are respectively divided according to the pore density based on the defect clustering statistical data.

[0034] Furthermore, in the present invention, random pores are quantitatively and precisely placed in the specifically divided density regions to generate a finite element model of three-dimensional braided ceramic matrix composites with randomly varying pore defects, and the damage failure mode is simulated and predicted with a subroutine highly matched with the material properties, and its mechanical properties are obtained accurately and quickly, taking into account the authenticity of the model and the calculation efficiency. Brief Description of the Drawings

[0035] Figure 1 It is the spatial structure of the three-dimensional braided unit cell yarn of the present invention;

[0036] Figure 2 It is the X-ray scan data of the present invention, and the obtained pore spatial distribution after filtering treatment;

[0037] Figure 3 It is the finite element RVE model of three-dimensional braided ceramic matrix composites of the present invention, where (a) is the finite element model of the composite material, (b) is the finite element model of the fiber, and (c) is the finite element model of the matrix;

[0038] Figure 4 It is the Monte Carlo random pore distribution of the present invention;

[0039] Figure 5The periodic boundary conditions and load conditions applied in the embodiments of the present invention, where (a) shows the application of the periodic boundary conditions and (b) shows the load application;

[0040] Figure 6 The finite element damage failure process of the present invention;

[0041] Figure 7 The comparison between the numerical results and the experimental stress-strain curve in the present invention. Detailed implementation manners

[0042] The following further describes the present invention in detail with reference to the accompanying drawings and specific embodiments:

[0043] The present invention discloses a damage prediction method for a three-dimensional braided ceramic matrix composite containing randomly varying pore defects, comprising the following steps:

[0044] Step (1): Perform cross-section scanning on the three-dimensional braided ceramic matrix composite obtained by chemical vapor deposition. The cross-section scanning method is X-ray scanning to obtain the mesoscopic internal morphology of the material. Adjust the observation threshold to eliminate noise interference, making the pore defects inside the material clearer and better distinguishing the defects and the matrix. The observation threshold includes the measured noise value, grayscale value, and image contrast. Preferably, after cutting, perform end-face scanning. Cut the three-dimensional braided ceramic matrix composite into tetrahedrons one by one for the convenience of subsequent steps.

[0045] The components of the three-dimensional braided ceramic matrix composite are a ceramic matrix. The ceramic matrix is formed after the preform (only carbon fiber) woven by fibers is impregnated with ceramic and then pyrolyzed, and it contains pore defects. The pores include the defects woven by fibers and the pore defects formed by the cured ceramic after impregnation. Separate the pores from other components by pre-data filtering and adjusting the grayscale threshold of the X-ray projection data, and pores of different sizes can be identified to obtain the pore distribution pattern.

[0046] Step 2: Perform clustering analysis on the internal pores of the material to obtain the internal pore parameters of the material and the defect distribution of the material;

[0047] Step 2.1: Randomly select n pore centroids as the basic centroids, and each basic centroid represents the pore position and size distribution;

[0048] Step 2.2: Measure each of the remaining pores. The pore measurement values include the position, size, and magnitude of the pores; calculate the Euclidean distance from the centroid of the remaining pores to each of the closest basic centroids, and classify them into the array where the centroid with the smallest mutual distance is located, and calculate the new centroid of each array.

[0049] The calculation method for the Euclidean distance of the selected centroid of the pore centroid position in three-dimensional space:

[0050]

[0051] Where x1, y1, and z1 are the spatial rectangular coordinates of the centroid, and x2, y2, and z2 are the spatial rectangular coordinates of the pores respectively.

[0052] Step 2.3: After all pore points are grouped, recalculate the positions of the centroids of each array according to the division situation to obtain n arrays; each array consists of pores that are closest to the centroid of the array before.

[0053] Step 2.4: Repeat Step 2.2 and Step 2.3 to calculate the distances from all pores to the newly calculated centroids, and re-divide all pores.

[0054] Step 2.5: Repeat Step 2.2 and Step 2.3 until the centroid does not change or reaches the specified maximum number of iterations, to obtain the internal pore parameters of the material and the defect distribution of the material. Use cluster analysis to obtain the centroid of the array set as the defect coordinates of the material. The pore parameters include the final centroid after pore clustering, i.e., the pore size and position.

[0055] Step 3: Establish a finite element mesoscopic model of the material according to the structural characteristics and dimensional features of the three-dimensional braided composite material. According to the pore parameters and the pore defect distribution of the material obtained in Step 2.5, establish a placement area in the finite element mesoscopic model, and this placement area is the above-mentioned pore area; the finite element mesoscopic model of the material contains eight fiber bundles with 4 spatial orientations. When dividing the pore placement area, the fiber bundles and the matrix are divided separately.

[0056] Step (4): Use the Monte Carlo algorithm to randomly generate pore units and input the set porosity into the pre-defined pore area; use the porosity of the array set obtained by cluster analysis as the porosity of the placement area.

[0057] Step (5): Assign material properties and apply periodic boundary conditions and loads; the braided material is a three-dimensional four-directional ceramic matrix composite obtained by chemical vapor deposition process, the braided matrix is C / SiC, and the braided yarn uses T800 - 12K braided yarn.

[0058] Step (6): Embed the damage analysis theory into the finite element model through the method of user-defined subroutine for iterative calculation; the damage analysis method of the matrix adopts the Christensen damage criterion, which can well predict the damage and yield characteristics of isotropic materials under various stress states; the fiber bundle damage criterion adopts the Hashin criterion, and the strength parameters adopt the Chamis mesoscopic strength model.

[0059] Step (7): Predict the failure situation of the material based on the calculation results of the finite element simulation software. The failure situation includes the damage degree in three directions of the fiber and the matrix, fiber debonding, shear failure of the matrix, etc.

[0060] Example

[0061] The present invention proposes a damage prediction method for a three-dimensional woven ceramic matrix composite containing randomly varying pore defects. Using the method idea for simulation, it includes the following steps:

[0062] (1) Cut the three-dimensional four-direction woven ceramic matrix composite obtained by chemical vapor deposition, perform tomography to obtain the mesoscopic internal morphology of the material, and adjust the observation threshold to obtain the internal pore defects of the material; as Figure 1 shown.

[0063] (2) Perform cluster analysis on the internal pores of the material to obtain the internal pore parameters of the material and the defect distribution of the material as shown in Figure 2 ; Through scanning analysis, the total porosity of the material is 2%, the number of pores is 8,850, and the pore volume range is from 0.0000089 mm 3 -0.3264498 mm 3 .

[0064] (2.1) Randomly select n pore centroids in the pore interval, and each centroid represents the pore position and size distribution;

[0065] (2.2) For each remaining pore measurement value, calculate its distance to the selected centroid and assign it to the array where the centroid with the smallest mutual distance is located. Calculate the centroid of each new array generated;

[0066] (2.3) After all pore points are grouped, recalculate the position of the centroid of each array according to the division situation, then iteratively calculate the distance from each sample point to the centroid of each array, and re-divide all sample points;

[0067] (2.4) Repeat (2.2) and (2.3) until the centroid does not change or reaches the specified maximum number of iterations to obtain the internal pore parameters of the material and the defect distribution of the material;

[0068] (3) Establish a finite element mesoscopic model of the material according to the structural characteristics and dimensional characteristics of the three-dimensional woven composite. The geometric size is 1.57 mm × 1.57 mm × 2.84 mm; the diameter of the fiber bundle is elliptical. Divide the pore predetermined placement area according to the pore distribution of the composite material, as Figure 3 ;

[0069] (4) Divide the model into periodic grids, randomly generate pore units using the Monte Carlo algorithm, and input the specified porosity into the pre-defined pore region. The pore size is a multiple of the grid size and is consistent with the statistical value, such as Figure 4 ;

[0070] (5) Assign material properties and apply periodic boundary conditions. Applying periodic conditions simulates the boundary conditions of the finite element meso-model in the actual macroscopic material, thus greatly reducing the finite element simulation calculation efficiency. In the simulation method, it is reflected as the displacement coupling relationship between the cube vertices of the finite element meso-model and the nodes on the boundary of the finite element model;

[0071] (6) Apply a tensile load to simulate the loading condition of the three-dimensional braided ceramic matrix composite, such as Figure 5 ;

[0072] (7) The damage analysis theory is embedded in the finite element model through the method of user-defined subroutines for iterative calculation. The damage analysis method for the matrix adopts the Christensen damage criterion, which can well predict the damage and yield characteristics of isotropic materials under various stress states; the damage criterion for the fiber bundle adopts the Hashin criterion, and the strength parameters adopt the Chamis meso-strength model;

[0073] (8) Predict the failure condition of the material according to the calculation results of the finite element simulation software, calculate the material damage condition, and respectively export the damage nephogram of the fiber and the damage nephogram of the matrix, such as Figure 6 . Comparing with the numerical results of the tensile test, it is found that the overall simulation results are in good agreement with the stress-strain curve obtained from the test, such as Figure 7 , indicating that the finite element model established based on this method can effectively predict the mechanical properties and damage conditions of the braided ceramic matrix composite.

[0074] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A damage prediction method for three-dimensional woven ceramic matrix composites containing random gradient pore defects, characterized in that: The following steps are involved: Step 1: Obtain the pore defects inside the composite material through cross-section scanning; Step 2, obtaining the pore parameters, defect distribution and porosity inside the composite material through cluster analysis; Step 3: Establish a finite element microscopic model and input pore parameters, defect distribution and porosity into the finite element microscopic model; Step 4: randomly generate pore elements for the finite element microscopic model using the Monte Carlo algorithm to obtain a finite element model with pores; Step 5: Assign material properties to the finite element model with pores and apply periodic boundary conditions and loads to obtain a finite element model with pores and boundary conditions and loads; Step 6: Using a damage analysis method to predict the finite element model with pores, boundary conditions, and loads, the damage and yield characteristics of the finite element model under various stress conditions are predicted; Step 7: Obtain the failure condition of the composite material based on the calculation results of damage and yield characteristics.

2. The damage prediction method for a three-dimensional woven ceramic matrix composite material containing random gradient pore defects according to claim 1, characterized in that: In step 1, the cross section is scanned by x-rays.

3. A damage prediction method for a three-dimensional woven ceramic matrix composite containing randomly varying pore defects according to claim 1, characterized in that, In step 1, the composite material is cut into tetrahedrons before scanning.

4. A damage prediction method for a three-dimensional woven ceramic matrix composite containing randomly varying pore defects according to claim 1, characterized in that, The specific process of step 2 is: 2.1 Randomly select n pore centroids as the base centroids; 2.2 Calculate the Euclidean distance between each remaining pore and the nearest base center, organize the Euclidean distances associated with each pore centroid into an array of the pore centroids, and calculate the new centroid of each array; 2.3 Calculate the Euclidean distances between all pores and their nearest new centroids, and divide the distance associated with each new centroid into an array of new centroids; 2.4 Repeat steps 2.2 and 2.3 to obtain the pore parameters and defect distribution inside the composite material until the center of mass no longer changes or the maximum number of iterations is reached. The center of mass of the array finally obtained is used as the defect coordinates of the composite material.

5. The damage prediction method for three-dimensional woven ceramic matrix composite materials containing random gradient pore defects according to claim 4, characterized in that: The calculation formula of the Euclidean distance is: Among them, x1, y1, z1 are the spatial rectangular coordinates of the center of mass, and x2, y2, z2 are the spatial rectangular coordinates of the pores.

6. A damage prediction method for a three-dimensional woven ceramic matrix composite containing randomly varying pore defects according to claim 1, characterized in that, In step 5, the material property is a three-dimensional four-directional ceramic matrix composite material, the braided matrix is C / SiC, and the braided yarn is T800-12K braided yarn.

7. A damage prediction method for a three-dimensional woven ceramic matrix composite containing randomly varying pore defects according to claim 1, characterized in that, In step 3, when dividing the predetermined delivery area of the pores, the fiber bundles and the matrix are divided separately.

8. A method for predicting damage of a three-dimensional woven ceramic matrix composite containing randomly varying pore defects according to claim 6, characterized in that, In step 6, the damage analysis method for the matrix is the Christensen damage criterion.

9. The damage prediction method for three-dimensional woven ceramic matrix composite materials containing random gradient pore defects according to claim 6, characterized in that: In step 6, the damage analysis method of the fiber bundle is the Hashin criterion, and the strength parameter adopts the Chamis micro-strength model.

10. The damage prediction method for three-dimensional woven ceramic matrix composite materials containing random gradient pore defects according to any one of claims 1 to 9, characterized in that: In step 7, the failure conditions include the degree of damage in three directions of the composite material, fiber debonding, and shear failure of the matrix.

Images