A method for constructing a microscopic conductive network of ceramic-based fiber bundle composite materials

Abstract

The present invention discloses a method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material. Three-dimensional image data of the ceramic-based fiber bundle composite material is obtained by XCT technology, and a deep learning model is used for training to obtain the three-dimensional spatial distribution results of the material components. Through volume analysis, the number and distribution pattern of single fibers without contact and multiple fibers in contact with each other are separated. The number and distribution pattern of single fibers without contact and multiple fibers in contact with each other are statistically obtained to distinguish whether there is contact between single fiber filaments. Non-contact single fiber filaments are regarded as individual resistors, and the mutual contact between multiple fibers is equivalent to a corresponding resistive conductive network according to the contact mode, and the matrix part is equivalent to a resistor. The present invention extracts the real structural data of the fibers in three-dimensional space, takes into account the characteristics of mutual contact between single fibers in the fiber bundle, and realizes the construction of a conductive network model of the ceramic-based fiber bundle composite material taking into account the real fiber contact characteristics.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent detection of material damage, and in particular relates to a method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material. Background Art

[0002] Braided ceramic matrix composites (CMCs) have become the material of choice in the aerospace industry due to their high specific strength, specific stiffness, high-temperature resistance, and designability. However, CMCs are often used under extreme conditions, resulting in complex damage patterns. To ensure the safety and reliability of CMCs during use, it is crucial to effectively monitor their damage behavior and assess their residual mechanical properties and service life.

[0003] Among the many intelligent damage detection technologies, the resistance method is a novel intelligent damage detection method that leverages the inherent properties of the material to monitor damage. When damage occurs within a material, its resistance changes accordingly, reflecting the close relationship between the material's mechanical properties and resistance. Therefore, by understanding the relationship between the material's mechanical properties and resistance, and by monitoring the resistance changes in real time, it is possible to accurately monitor the material's damage extent and remaining lifespan. As a fundamental component of woven ceramic matrix composites, the mechanical resistance properties of ceramic-based fiber bundle composites (CMFBs) are crucial to the overall material performance. These mechanical resistance properties are closely related to their microscopic conductive network structure. Establishing an accurate microscopic conductive network is fundamental for damage identification and mechanical property evaluation. However, existing technologies often ignore the actual contact between fibers, simplifying the fibers to a uniform distribution and constructing an equivalent resistance conductive network. This simplification results in inaccurate equivalent resistance network models, making it difficult to accurately reflect the actual relationship between material damage and resistance change. Furthermore, fiber bundles are composed of hundreds to thousands of individual fiber filaments, and the contact characteristics between these filaments (such as position, length, and number) exhibit a complex three-dimensional distribution. Traditional manual methods for acquiring these three-dimensional features are both time-consuming and costly. The complex conductive paths formed by fiber contact make the conductive network inside the fiber bundle extremely large and complex. It is a huge challenge to effectively describe these contact characteristics and establish an accurate equivalent resistance network model.

[0004] In order to solve the above problems and obtain the relationship between the mechanical resistance properties and the microscopic conductive network of ceramic fiber bundle composites, a method for constructing a microscopic conductive network of ceramic-based fiber bundle composites based on real contact is urgently needed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material in response to the above-mentioned deficiencies in the prior art.

[0006] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0007] A method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material comprises the following steps:

[0008] Step 1: Obtain XCT three-dimensional image data of ceramic-based fiber bundle composites;

[0009] Step 2: After cropping, any part of the XCT three-dimensional image data is obtained as a training set for deep learning model training. The components of the ceramic-based fiber bundle composite material in the training set are manually calibrated as the target set.

[0010] Step 3: Based on the training set and target set of the deep learning model, set the parameters of the deep learning process and train the component segmentation model of the ceramic-based fiber bundle composite material.

[0011] Step 4: Use the ceramic-based fiber bundle composite component segmentation model to identify and segment the complete XCT 3D image data of the ceramic-based fiber bundle composite to obtain the 3D spatial distribution results of the fiber single filaments, interfaces, matrix, and pores;

[0012] Step 5: Based on the three-dimensional spatial distribution results of the fiber monofilaments, the number and distribution patterns of single fibers without contact and multiple fibers in contact with each other are separated through volume analysis. When the fiber monofilaments are adjacent and without contact, different grayscale values ​​are assigned to the fiber monofilaments. When there is contact between the fiber monofilaments, the contacting fiber monofilaments are assigned the same grayscale value, thereby obtaining the distribution results of whether the fiber monofilaments are in contact with each other.

[0013] Step 6: Count the number and distribution of single fibers without contact and multiple fibers in contact with each other, distinguish whether there is contact between single fibers, separate the results of no contact between fibers, and treat the fibers as resistors. When single fibers are not in contact, the current only flows along the direction of the fibers, that is, when the fibers are not in contact, there is only a parallel relationship between the resistors to form a conductive network of resistors.

[0014] Step 7: The mutual contact between multiple fibers is based on the arrangement and combination of two fibers in contact with each other. The fiber contact types are counted based on the actual contact situation to obtain the resistance conductive network formed by the mutual contact between multiple fibers.

[0015] Step 8: The matrix part of the ceramic-based composite material is equivalent to a resistor. Due to the existence of the non-conductive interface, the resistance conductive network formed by the current conducting along the matrix and the current conducting along the fiber is in a parallel relationship. After the fibers are non-contact, the fibers are in contact with each other, and the matrix conduction is equivalent to a resistor, a resistance conductive network is constructed to realize the construction of a microscopic conductive network of the ceramic-based fiber bundle composite material.

[0016] To optimize the above technical solutions, specific measures taken also include:

[0017] In step 1, XCT three-dimensional image data of the ceramic-based fiber bundle composite material is obtained by Micro-CT.

[0018] In step 2, the ceramic-based fiber bundle composite material components include three types of components: matrix, fiber, interface and pores.

[0019] In step 3, deep learning parameters include the deep learning network type, the ratio of training set to validation set data, model learning rate, and number of training times.

[0020] The present invention has the following advantages:

[0021] 1. The present invention combines the Micro-CT three-dimensional reconstruction model of the microstructure of ceramic-based fiber bundle composite materials with deep learning to obtain a deep learning model for the rapid segmentation of ceramic-based fiber bundle composite materials, avoiding the repetitive work of manually calibrating the components within the fiber bundle and achieving accurate and rapid segmentation of the various components of the fiber bundle.

[0022] 2. The present invention extracts the real structural data of the fiber in three-dimensional space, takes into account the contact characteristics between individual fibers within the fiber bundle, and realizes the construction of a conductive network model of ceramic-based fiber bundle composite materials that takes into account the real fiber contact characteristics.

[0023] 3. The method provided by the present invention is not only applicable to the segmentation of components of unidirectional ceramic-based fiber bundle composite materials and the construction of microscopic conductive networks, but is also applicable to the construction of conductive networks of various composite materials with conductive capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 Schematic diagram of the training set and target set in the deep learning process of the present invention; the left figure is Figure 1 (a) is a diagram of the training set; the right diagram is Figure 1 (b) is a schematic diagram of the target set;

[0025] Figure 2 It is an iterative curve graph of the accuracy of the deep learning training segmentation model of the present invention;

[0026] Figure 3 A three-dimensional reconstruction of the ceramic-based fiber bundle composite material used in the present invention;

[0027] Figure 4 Graph showing the segmentation results of each component obtained by the segmentation model of the present invention;

[0028] Figure 5 This is a three-dimensional spatial distribution diagram of each component of the ceramic-based fiber bundle composite material of the present invention;

[0029] Figure 6 This is a diagram showing the results of contact screening of the fiber monofilament of the present invention;

[0030] Figure 7 This is a three-dimensional result diagram of a single fiber without contact in the present invention;

[0031] Figure 8 Schematic diagram of fiber contact type and equivalent contact; the left figure is Figure 8 (a) is the basic type of two fibers contacting each other. The right figure is Figure 8 (b) Schematic diagram of the equivalent resistance of three contact methods.

[0032] Figure 9 Schematic diagram of the microscopic conductive network of the ceramic fiber bundle composite material of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely used to explain this application and are not intended to limit this application. Based on the embodiments provided in this application, all other embodiments obtained by those of ordinary skill in the art without making any creative efforts are within the scope of protection of this application.

[0034] Obviously, the drawings described below are merely examples or embodiments of the present application. Those skilled in the art can, without inventive effort, apply the present application to other similar scenarios based on these drawings. Furthermore, it is also understood that, although the effort involved in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, changes in design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as an insufficiency of the content disclosed in this application.

[0035] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0036] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "a", "an", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or units (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The words "multiple" / "several" used in this application refer to two or more. "And / or" describes the association relationship between associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, or B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.

[0037] A method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material comprises the following steps:

[0038] Step 1: First, obtain XCT three-dimensional image data of ceramic-based fiber bundle composites based on Micro-CT.

[0039] Step 2: Based on the XCT 3D image data obtained in step 1, select part of the XCT 3D image data for deep learning to build a model for identifying and classifying the components of ceramic-based fiber bundle composite materials. The training set for deep learning is as follows: Figure 1 As shown in (a), Figure 1 (b) shows the target set of each component of the ceramic-based fiber bundle composite, which was manually calibrated, including matrix, fiber, interface, and pore. These four categories constitute the complete classification of the ceramic-based fiber bundle composite XCT 3D image data.

[0040] Step 3: Based on the above deep learning training set and target set, set the relevant parameters in the deep learning process. The network type selected for deep learning is BackbonedNet, the data volume of the training set and the validation set is 3:1, the model learning rate is 0.001, the number of training times is 100,000 times, etc., and the ceramic matrix fiber bundle composite component segmentation model - CMCs_SegmentationModel is obtained through training. Figure 2The figure shows the changes in the recognition accuracy of the model on the training set and validation set and the loss function during the training process. After 100,000 iterative training, the recognition accuracy of the segmentation model for each component is close to 1.

[0041] Step 4: The segmentation model obtained by deep learning - CMCs_SegmentationModel Figure 3 The XCT three-dimensional image data of the ceramic-based fiber bundle composite material shown in FIG is subjected to component identification and segmentation processing to obtain the following Figure 4 The results shown in the figure were separated as Figure 5 The three-dimensional spatial distribution results of fibers, interfaces, matrix and pores are shown.

[0042] Step 5: Based on the three-dimensional spatial distribution results of the fiber monofilaments obtained in step 4, analyze whether the fibers are in contact by volume analysis. Figure 6 As shown, the analysis results are represented by grayscale values: non-contacting fiber filaments are assigned different grayscale values, while contacting fiber filaments are assigned the same grayscale value. The results show a total of 144 grayscale values ​​for the fiber filaments, indicating whether or not contact between the fiber filaments is divided into 144 separate sections. The detailed distribution of the fiber filaments is shown in the table below.

[0043] Fiber contact category quantity Fiber contact category quantity Fiber contact category quantity Fiber-free contact 85 7 fibers in contact 1 19 fibers in contact 1 2 fibers in contact 18 9 fibers in contact 1 37 fibers in contact 1 3 fibers in contact 16 10 fibers in contact 1 50 fibers in contact 1 4 fibers in contact 8 14 fibers in contact 1 74 fibers in contact 1 5 fibers in contact 5 16 fibers in contact 1 Fiber type quantity 144 6 fibers in contact 2 18 fibers in contact 1 Total fiber count 492

[0044] Step 6: Based on the number and distribution of single fibers without contact and multiple fibers in contact with each other obtained in step 5, the number of single fibers without contact is 85, such as Figure 7 Shown are independent, non-contacting fiber filaments. These non-contacting fibers conduct electricity only along the fiber direction. The fibers can be compared to resistors. In this case, these non-contacting fiber filaments form a resistive network conductive network formed by parallel connection of individual fiber resistors.

[0045] Step 7: The contact between multiple fibers is based on the arrangement and combination of two fibers in contact with each other. Based on the actual contact situation, the basic types of contact between two fibers are as follows: Figure 8 As shown in (a), there are three situations: contact at one end, contact in the middle, and complete contact (equivalent to a fiber); Figure 8 (b) Schematic diagram of the equivalent resistance of three contact modes.

[0046] Step 8: In this example, the interface component of the ceramic-based fiber bundle composite material is non-conductive boron nitride, so the conductive components mainly consist of two parts: the matrix and the fiber. The matrix conductivity and the fiber conductivity are considered to be in parallel, that is, the matrix part of the ceramic-based fiber bundle composite material is equivalent to a resistor in parallel with the fiber conductive network. Through the above-mentioned fiber contact method, a structure such as Figure 9 The microscopic conductive network of the ceramic-based fiber bundle composite material is shown. The microscopic conductive network of the ceramic-based fiber bundle composite material is constructed based on three-dimensional reconstruction.

[0047] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material, characterized by: The following steps are involved: Step 1: Obtain XCT three-dimensional image data of ceramic-based fiber bundle composites; Step 2: After cropping, any part of the XCT three-dimensional image data is obtained as a training set for deep learning model training. The components of the ceramic-based fiber bundle composite material in the training set are manually calibrated as the target set. Step 3: Based on the training set and target set of the deep learning model, set the parameters of the deep learning process and train the component segmentation model of the ceramic-based fiber bundle composite material. Step 4: Use the ceramic-based fiber bundle composite component segmentation model to identify and segment the complete XCT 3D image data of the ceramic-based fiber bundle composite to obtain the 3D spatial distribution results of the fiber single filaments, interfaces, matrix, and pores; Step 5: Based on the three-dimensional spatial distribution results of the fiber monofilaments, the number and distribution patterns of single fibers without contact and multiple fibers in contact with each other are separated through volume analysis. When the fiber monofilaments are adjacent and without contact, different grayscale values ​​are assigned to the fiber monofilaments. When there is contact between the fiber monofilaments, the contacting fiber monofilaments are assigned the same grayscale value, thereby obtaining the distribution results of whether the fiber monofilaments are in contact with each other. Step 6: Count the number and distribution of single fibers without contact and multiple fibers in contact with each other, distinguish whether there is contact between single fibers, separate the results of no contact between fibers, and treat the fibers as resistors. When single fibers are not in contact, the current only flows along the direction of the fibers, that is, when the fibers are not in contact, there is only a parallel relationship between the resistors to form a conductive network of resistors. Step 7: The mutual contact between multiple fibers is based on the arrangement and combination of two fibers in contact with each other. The fiber contact types are counted based on the actual contact situation to obtain the resistance conductive network formed by the mutual contact between multiple fibers. Step 8: The matrix part of the ceramic-based composite material is equivalent to a resistor. Due to the existence of the non-conductive interface, the resistance conductive network formed by the current conducting along the matrix and the current conducting along the fiber is in a parallel relationship. After the fibers are non-contact, the fibers are in contact with each other, and the matrix conduction is equivalent to a resistor, a resistance conductive network is constructed to realize the construction of a microscopic conductive network of the ceramic-based fiber bundle composite material.

2. The method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material according to claim 1, wherein: In step 1, XCT three-dimensional image data of the ceramic-based fiber bundle composite material is obtained by Micro-CT.

3. The method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material according to claim 1, wherein: In step 2, the ceramic-based fiber bundle composite material components include four categories: matrix, fiber, interface and pores.

4. The method for constructing a microscopic conductive network of a ceramic-based fiber bundle composite material according to claim 1, wherein: In step 3, deep learning parameters include the deep learning network type, the ratio of training set to validation set data, model learning rate, and number of training times.

Images