A method for establishing a high-fidelity unit cell model of carbon fiber fabric

Abstract

The present invention relates to a method for establishing a high-fidelity unit cell model of a carbon fiber fabric. The smallest non-repeating fabric unit cell structure is selected according to the fiber fabric used, and the thickness of the fiber fabric, the cross-sectional area of the fiber bundle, and the fabric weaving pattern are obtained from the optical image; the fiber bundle is discretized into a specified number of representative fiber filaments using an equivalent volume, and the representative fiber filaments are woven in the same form according to the fiber fabric weaving pattern; appropriate periodic boundary conditions are applied to the model, and dynamic simulation is carried out in combination with the specified working conditions; the fiber bundle is reconstructed according to the distribution state of the representative fiber filaments after simulation; the interference part of the fiber bundle is eliminated, and finally a high-fidelity unit cell RVE model under the specified working conditions is obtained.

Description

Technical Field

[0001] The present invention belongs to the field of simulation modeling of resin-based carbon fiber reinforced composites, and relates to a method for establishing a high-fidelity unit cell model of carbon fiber fabric. Background Art

[0002] Resin-based carbon fiber reinforced composites have the advantages of high specific strength, high specific modulus, good fatigue resistance, corrosion resistance, good designability, etc. Their characteristics of light weight and high strength give them unique application advantages in advanced industrial manufacturing fields such as aerospace, vehicles, and ships.

[0003] Among them, carbon fiber fabric is used as the reinforcing phase and is combined with resin by corresponding forming techniques to manufacture composite products. During the manufacturing process, two-dimensional carbon fiber fabric is usually first made into a fiber preform with a specific designed shape, and then resin infiltration and curing and subsequent processes are carried out to finally complete the product manufacturing. The flow of resin in the fiber preform is affected by the structure of the fiber preform, and the corresponding flow process will affect the generation of pore defects during the mold filling process. Therefore, establishing a unit cell model of the fiber fabric under specified working conditions is crucial for the prediction and control of related processes.

[0004] To save manufacturing costs and reduce the manufacturing cycle, it is very valuable to carry out numerical simulation before determining the process plan to determine process parameters and optimize the process path. Due to the periodic and repetitive structural characteristics of the fiber fabric, the related analysis often uses the fabric unit cell RVE (representative volume element) model. However, most fabric RVE models are modeled by combining fiber bundles with a constant geometric shape cross-section and fabric weaving methods, which has a large gap with the real structure of the fabric under actual working conditions, resulting in inaccurate related analysis based on the model and insufficient application value of the corresponding numerical simulation in actual engineering. How to establish fabric RVE models under different specified working conditions is the premise for accurately predicting and analyzing the seepage behavior of resin in the preform. At the same time, establishing a corresponding high-fidelity RVE model according to the macro-micro structure of the fabric is a necessary condition for multi-scale mechanical property analysis of the fabric.

[0005] The present invention proposes a method for establishing a high-fidelity unit cell RVE model of two-dimensional carbon fiber fabric. First, the fiber bundles in the fabric unit cell are discretized into a number of representative fiber filaments, and dynamic compaction simulation is carried out on these discrete fiber filaments to obtain the fiber filament distribution. Then, corresponding cross-section reconstruction and fiber bundle generation are carried out on the fiber bundles to which the discrete fiber filaments belong, and interference between fiber bundle models is eliminated and gap control is carried out to complete the construction of a high-fidelity fabric unit cell RVE model under specified working conditions. Summary of the Invention

[0006] Technical Problems to be Solved

[0007] To avoid the deficiencies of the prior art, the present invention proposes a method for establishing a high-fidelity unit cell model of carbon fiber fabric, providing a method for establishing a high-fidelity unit cell RVE model of two-dimensional carbon fiber fabric.

[0008] Technical solution

[0009] A method for establishing a high-fidelity unit cell model of carbon fiber fabric, characterized in that the steps are as follows:

[0010] Step 1: Process the digital image of the fiber fabric to obtain data such as fabric thickness, major and minor axes and area of the cross-section of warp and weft yarn fiber bundles, and porosity;

[0011] Step 2: According to the cross-sectional areas of the warp and weft yarn fiber bundles and the porosity, discretize the warp and weft yarn fiber bundles into a number of representative fiber filaments, where the radius of the representative virtual fiber filament is:

[0012]

[0013] S f =π·r f 2

[0014] where S is the cross-sectional area of a single bundle of fibers, S f is the cross-sectional area of a single discrete fiber, N is the number of discrete fibers, r f is the radius of the representative virtual fiber filament, is the volume fraction of fiber filaments in the fiber bundle;

[0015] Step 3: According to the weaving pattern of the fabric and the unit cell size of the fabric, establish the spatial coordinates of the control nodes of the centerlines of the discrete representative virtual fiber filaments along the paths of the warp and weft yarn bundles of the fabric, and perform spline curve interpolation and radius assignment to complete the construction of the three-dimensional model of the discrete fibers;

[0016] Step 4: Import the established three-dimensional model of the discrete fiber filaments into the dynamics simulation software. The model is placed between two rigid plates, and each fiber filament is established as a digital chain connected without pins by a number of beam elements, and the periodic boundary conditions are assigned according to the following formula:

[0017] Uai = Ubi, i = 1, 2, 3, 4, 5, 6

[0018] where a and b represent the end points of the fiber bundle connected by the line, U represents the degree of freedom, and i represents the degree of freedom direction of the corresponding end point;

[0019] Then fix the rigid lower plate, apply a specified displacement to the rigid upper plate, control the thickness of the discrete fiber filaments after compaction through the relative displacement of the two plates, and apply a load condition of one end fixed and the other end applied with a concentrated force of appropriate magnitude to the discrete fiber filaments during this process;

[0020] Step 5: First, export and reconstruct the discrete fiber filaments after compaction simulation, and summarize the representative fiber filaments belonging to the same warp or weft yarn into a set, which is called a fiber bundle;

[0021] Divide each fiber bundle into several cross-sections along the axial direction of the fiber bundle center line. The cross-section is perpendicular to the fiber center line at that place, and obtain the corresponding intersection points of the representative fiber filaments on the cross-section;

[0022] Use a rope winding algorithm based on the force offset principle to perform height reconstruction on the cross-section of the fiber bundle with concave and convex features;

[0023] Step 6: Combine the warp and weft yarn fiber bundle models to construct a fabric unit cell model according to the weaving method of the selected fabric, and use an interference elimination algorithm to achieve interference elimination and gap insertion in the tiny area of the model, and finally complete the reconstruction of the fabric unit cell model.

[0024] The rope winding algorithm based on the force offset principle is as follows: First, establish a closed flexible rope that wraps the fiber filament set, and discretize the rope into a chain of spheres composed of a series of spheres with a radius of r; set virtual tensile and bending springs between the spheres and endow them with corresponding spring characteristics to maintain the smoothness and continuity of the sphere chain during movement; apply a centripetal force to the flexible rope to make the discrete small balls on it move towards the center of the area surrounded by the flexible rope at the same time; at the same time, set the repulsive force related to the overlap degree between the outer small balls and the fibers; during the movement, calculate the resultant force received by all small balls, including the tensile spring force, bending spring force, repulsive force and centripetal force, and configure a small displacement in the same direction for it according to the resultant force it receives. Finally, when the resultant force of all small balls reaches equilibrium, stop the movement; when the distance between the centers of adjacent discrete spheres is less than the radius r of the sphere, perform overlap deletion, and when the distance between adjacent centers is greater than 2r, perform sphere interpolation to supplement. Finally, wrap the discrete fiber filaments in the same bundle. When the whole sphere chain reaches the equilibrium state with the minimum potential energy, stop the movement of the spheres and output the fiber cross-section contour; perform surface interpolation on the fiber bundle contour to reconstruct the complete warp and weft yarn fiber bundles in the fabric unit cell.

[0025] The process of the interference elimination algorithm is as follows: Discretize the surface of the fiber bundle into a dense point cloud set, define the in-plane space region as Ω, and at the same time define the surface of a fiber bundle as a zero potential energy surface Φ0. Construct an Euclidean distance field. Define the potential energy as negative for the points inside the fiber bundle, and positive otherwise. The potential energy size is related to the distance d between the point p and the zero potential energy surface, that is:

[0026] Φ = -kd, p ∈ Ω

[0027]

[0028] First, for the point with negative potential energy, the minimum distance from the zero-potential energy surface and the unit normal vector corresponding to the corresponding point on the zero-potential energy surface are perpendicular to the outside of the surface. Solve to make this point move along When all the points with negative potential energy are moved to make the potential energy zero, the interference between this fiber bundle and other fiber bundles is eliminated; then, judge the potential energy of the points with non-zero potential energy. When the potential energy is greater than the set value Φ t Keep the position, otherwise move this point along the normal direction corresponding to the zero-potential energy surface of the fiber it belongs to by a distance s until the potential energy is greater than the set value Φ t , so as to realize the insertion and control of the fiber bundle gap, and the minimum gap is d min , that is:

[0029]

[0030] Φ t = k * d min .

[0031] The spring characteristic is the restoring force of a tensile or bending spring.

[0032] Beneficial effects

[0033] A method for establishing a high-fidelity unit cell model of a carbon fiber fabric proposed by the present invention selects the smallest non-repeating fabric unit cell structure according to the fiber fabric used, and obtains the fiber fabric thickness, the cross-sectional area of the fiber bundle, and the fabric weaving method according to the optical image; uses the equivalent volume to discretize the fiber bundle into a specified number of representative fiber filaments, and weaves the representative fiber filaments in the same form according to the fiber fabric weaving method; applies appropriate periodic boundary conditions to this model and conducts a dynamic simulation in combination with the specified working conditions; reconstructs the fiber bundle according to the distribution state of the representative fiber filaments after the simulation; eliminates the interference of the fiber bundle interference part, and finally obtains a high-fidelity unit cell RVE model under the specified working conditions. Description of the drawings

[0034] Figure 1 : Satin two-dimensional plane fabric sample

[0035] Figure 2 : Discrete representative fiber filaments

[0036] Figure 3 : Discrete fiber model compaction simulation process

[0037] (a) Discrete fibers and the force field applied to them;

[0038] (b) Method for applying periodic boundary conditions

[0039] (c) Fiber compaction process

[0040] (d) Fiber compaction result

[0041] Figure 4 : Initial state of the rope winding algorithm

[0042] Figure 5 : Obtain the cross-sectional profile of the fiber bundle by the rope winding algorithm

[0043] Figure 6 : Perform multi-section surface interpolation on the cross-sectional profile of the fiber bundle

[0044] Figure 7 : Interference and elimination between fiber bundles

[0045] (a) Interference between fiber bundles

[0046] (b) Interference elimination and gap control

[0047] Figure 8 ; Unit cell model of satin fabric Specific implementation manner

[0048] The present invention will be further described in conjunction with embodiments and the accompanying drawings as follows:

[0049] The specific steps included in this method are as follows:

[0050] Step 1: Acquire an image of the fiber fabric, which contains information such as the weaving pattern of the fiber fabric, the fabric thickness, the cross-sections of the warp and weft yarns of the fabric, etc. Use image processing software to process the image in combination with a size calibration piece to obtain parameters such as the fabric thickness, the major and minor axes and area of the cross-sections of the warp and weft fiber bundles, and the porosity.

[0051] Step 2: Discretize the warp and weft fiber bundles into a number of representative virtual fiber filaments according to the cross-sectional areas of the warp and weft fiber bundles and the volume fraction of the fiber filaments in the bundle. The radius of the representative virtual fiber filament is calculated according to the following formula:

[0052]

[0053] S f = π·r f 2

[0054] where S is the cross-sectional area of a single fiber bundle, S f is the cross-sectional area of a single discrete fiber, N is the number of discrete fibers, r f is the radius of the representative virtual fiber filament, is the volume fraction of the fiber filaments in the fiber bundle.

[0055] Step 3: Combine the weaving pattern of the used fabric and the unit cell size of the fabric, establish the spatial coordinates of the control nodes of the centerlines of the discrete representative virtual fiber filaments along the paths of the warp and weft yarn bundles of the fabric, and perform spline curve interpolation and radius assignment to complete the construction of the three-dimensional model of the discrete fibers.

[0056] Step 4: Import the established three-dimensional model of discrete fiber filaments into the dynamics simulation software. The model is placed between two rigid plates, and each fiber filament is established as a digital chain connected without pins by a number of beam elements. And the periodic boundary conditions shown in Figure 3 (b) are assigned according to the following formula.

[0057] Uai = Ubi, i = 1, 2, 3, 4, 5, 6

[0058] where a and b represent the endpoints of the fiber bundle indicating the connection line in the figure, U represents the degree of freedom, and i represents the degree of freedom direction of the corresponding endpoint.

[0059] Then fix the rigid lower plate, apply a specified displacement to the rigid upper plate, control the thickness after compaction of the discrete fiber filaments through the relative displacement of the two plates, and apply a load condition where one end is fixed and a concentrated force of appropriate magnitude is applied to one end to the discrete fiber filaments during this process.

[0060] Step 5: First, export and reconstruct the discrete fiber filaments after the compaction simulation, and summarize the representative fiber filaments belonging to the same warp or weft yarn into a set, which is called a fiber bundle. Each fiber bundle is divided into several cross-sections along the axial direction of the fiber bundle center line. The cross-sections are perpendicular to the fiber center line at that place, and the corresponding intersection points of the representative fiber filaments on the cross-sections are obtained. To obtain the cross-section of the discrete fiber model and solve the problem of poor curvature continuity of the traditional convex hull algorithm, the height of the cross-section of the fiber bundle with concave and convex features is reconstructed, and a rope winding algorithm based on the force offset principle is used. The main implementation method of this algorithm is as follows: First, establish a closed flexible rope (wrapping the fiber filament set), and discretize the rope into a chain of spheres composed of a series of spheres with a radius of r. Virtual tensile and bending springs are set between the spheres, and their corresponding spring characteristics (tensile and bending spring restoring forces) are given to maintain the smoothness and continuity of the sphere chain during movement. Among them, both the tensile spring restoring force and the bending spring force are related to the deformation of the spring. Apply a centripetal force to the flexible rope to make the discrete small balls on it move towards the center of the area surrounded by the flexible rope at the same time. At the same time, set the repulsive force between the outer small balls and the fibers (related to the overlap degree) to ensure that there is as little overlap between the small balls and the fibers as possible during the centripetal movement. During the movement, calculate the resultant force received by all the small balls, including the tensile spring force, the bending spring force, the repulsive force and the centripetal force. Configure a small displacement in the same direction for it according to the resultant force it receives. Finally, stop the movement when the resultant force of all the small balls reaches equilibrium. When the distance between the centers of adjacent discrete spheres is less than the radius r of the sphere, overlapping deletion is performed. When the distance between adjacent centers is greater than 2r, sphere interpolation is performed to supplement. Finally, the discrete fiber filaments in the same bundle are wrapped. When the sphere chain as a whole reaches the equilibrium state with the minimum potential energy, stop the movement of the spheres and output the fiber cross-section contour. Perform surface interpolation on the fiber bundle contour to reconstruct the complete warp and weft yarn fiber bundles in the fabric unit cell.

[0061] Step 6: Combine the warp and weft yarn fiber bundle models to construct a fabric unit cell model according to the weaving pattern of the selected fabric. During this process, there may be minor interference in the micro regions between the warp and weft yarn fiber bundles, which is a problem introduced due to the inherent limitations of dynamic numerical simulation. To enable this model to be more widely applied in numerical simulation prediction, an interference elimination algorithm is used to eliminate the minor interference in the model and insert gaps, and finally complete the reconstruction of the fabric unit cell model. The main process of this algorithm is as follows: Discretize the surface of the fiber bundle into a dense point cloud set, define the in-plane space region as Ω, and at the same time define the surface of a fiber bundle as the zero potential surface Φ0. Construct the Euclidean distance field. For points inside this fiber bundle, define the potential energy as negative, and vice versa as positive. The magnitude of the potential energy is related to the distance d between the point (p represents the point) and the zero potential surface, that is:

[0062] Φ = -kd, p ∈ Ω

[0063]

[0064] First, solve for the minimum distance between the point with negative potential energy and the zero potential surface and the unit normal vector corresponding to the corresponding point on the zero potential surface (defined as perpendicular to the surface and outward), and move the point along . When all the points with negative potential energy are moved to make the potential energy zero, the interference between this fiber bundle and other fiber bundles is eliminated. Then, judge the potential energy of the points with non-zero potential energy. When the potential energy is greater than the set value Φ t , keep the position, otherwise move the point along the normal direction corresponding to the zero potential surface of the fiber it belongs to by a distance s until the potential energy is greater than the set value Φ t , thus realizing the insertion and control of the gaps between fiber bundles, and the minimum gap is d min , that is:

[0065]

[0066] Φ t = k*d min .

[0067] Specific embodiment: Taking the satin fiber fabric shown in Figure 1 as an example, this satin pattern is a five - weft satin. Combining with the attached drawings, a method for establishing a high - fidelity unit cell model of a two - dimensional planar fabric is described.

[0068] 1. Select a 4 cm × 4 cm area containing the fabric unit cell from the satin fabric to prepare a sample. Place the single - layer satin fabric between two transparent acrylic plates, control the thickness between the plates through gaskets, and use bolts to tighten and fix the thickness.

[0069] 2. Use a camera combined with dimensional calibration to record the weaving pattern of satin fabrics, and include information such as fabric thickness, cross-sections of warp and weft yarns of the fabric, etc.

[0070] 3. Use image processing software to process the images, and calibrate parameters such as fabric thickness, major and minor axes and areas of cross-sections of warp and weft fiber bundles, volume fraction of fiber filaments in the bundle, etc.

[0071] 4. According to the cross-sectional area S of the warp and weft fiber bundles and the volume fraction of fiber filaments in the bundle Discretize the warp and weft fiber bundles into 61 representative fiber filaments as shown in Figure 2 . The radius of the fiber filament is calculated according to the following formula:

[0072]

[0073] S f = π·r f 2

[0074] where S is the cross-sectional area of a single fiber bundle, is the volume fraction of fiber filaments in the fiber bundle, S f is the cross-sectional area of a single discrete fiber, and r f is the radius of the representative fiber filament.

[0075] 5. Combine the weaving pattern of the used fabric to establish the spatial coordinates and paths of the control nodes of the centerlines of the discrete representative fiber filaments, and perform spline curve interpolation and radius assignment to complete the construction of the three-dimensional model of the discrete fibers shown in Figure 2 .

[0076] 6. Import the established three-dimensional model of the discrete fiber filaments into the dynamics simulation software. The model is placed between two rigid plates as shown in Figure 3 (c), and each fiber filament is established as a digital chain with pinless connections by several beam elements, and the periodic boundary conditions shown in Figure 3 (b) are assigned according to the following formula.

[0077] Uai = Ubi, i = 1, 2, 3, 4, 5, 6

[0078] where a and b represent the end points of the fiber bundle connected by the line in Figure 3 (b), U represents the degree of freedom, and i represents the direction of the degree of freedom of the corresponding end point.

[0079] Then fix the rigid lower plate, apply a specified displacement to the rigid upper plate, control the thickness of the discrete fiber filaments after compaction through the relative displacement of the two rigid plates, and apply a load condition with one end fixed and a concentrated force of appropriate magnitude applied to one end to the discrete fiber filaments during this process, as shown in Figure 3(a) As shown, the fiber density and elastic modulus are determined according to the actual fabric material parameters. To improve the calculation efficiency, mass scaling can be defined when the simulation results do not change significantly. The fiber mass is magnified by 10 times, and the results are as shown in Figure 3 (d).

[0080] 7. Export and reconstruct the discrete fiber filaments after compaction simulation, and summarize the representative fiber filaments belonging to the same warp or weft yarn into a set, which is called a fiber bundle. Each fiber bundle is divided into several cross-sections along the axial direction of the fiber bundle center line. The cross-section is perpendicular to the fiber center line at that place, and the corresponding intersection points of the representative fiber filaments on the cross-section are obtained. As shown in Figure 4 、 5 , the winding rope algorithm is used to obtain the cross-section contour of the warp and weft yarn fiber bundles.

[0081] 8. As shown in Figure 6 , the obtained cross-section contour of the fiber bundle is interpolated by multi-section surfaces to reconstruct the fiber bundle.

[0082] 9. As shown in Figure 7 , the warp and weft yarn fiber bundles are combined, and the fiber bundle interference elimination algorithm is used to eliminate the small interference of the fiber bundles and insert the specified gap. Finally, the unit cell model of the satin two-dimensional plane fabric shown in Figure 8 is established.

Claims

1. A method for establishing a high-fidelity unit cell model of a carbon fiber fabric, characterized in that The steps are as follows: Step 1: Process the digital image of the fiber fabric to obtain data on fabric thickness, the major and minor axes and area of the cross-section of warp and weft yarn fiber bundles, and porosity; Step 2: Discretize the warp and weft yarn fiber bundles into a number of representative fiber filaments according to the cross-sectional area of the warp and weft yarn fiber bundles and porosity, where the radius of the representative fiber filament is: S f ＝π·r f 2 Its S is the cross-sectional area of a single fiber bundle, S f is the cross-sectional area of a single discrete fiber, N is the number of discrete fibers, r f is the radius of a representative fiber filament, is the volume fraction of fiber filaments within the fiber bundle; Step 3: According to the weaving pattern of the fabric and the fabric unit cell size, establish the spatial coordinates of the control nodes of the centerlines of the discrete representative fiber filaments along the paths of the warp and weft yarn bundles of the fabric, and perform spline curve interpolation and radius assignment to complete the construction of the discrete fiber three-dimensional model; Step 4: Import the established discrete fiber three-dimensional model into the dynamics simulation software. The model is placed between two rigid plates, and each fiber filament is established as a digital chain connected without pins by a number of beam elements, and periodic boundary conditions are assigned according to the following formula: Uai = Ubi, i = 1, 2, 3, 4, 5, 6 where a and b represent the end points of the connected fiber bundles, U represents the degree of freedom, and i represents the direction of the corresponding end point degree of freedom; Then fix the rigid lower plate, apply a specified displacement to the rigid upper plate, control the thickness after compaction of the discrete fiber filaments through the relative displacement of the two plates, and during this process, assign a load condition of fixed at one end and applied with a concentrated force of appropriate magnitude at the other end to the discrete fiber filaments; Step 5: First, export and reconstruct the discrete fiber filaments after the compaction simulation, and summarize the representative fiber filaments belonging to the same warp or weft yarn bundle into a set, which is called a fiber bundle; Divide each fiber bundle into a number of cross-sections along the axial direction of the fiber bundle centerline. The cross-section is perpendicular to the centerline of the fiber bundle, and obtain the corresponding intersection points of the representative fiber filaments on the cross-section; Use a winding rope algorithm based on the force offset principle to reconstruct the height of the fiber bundle cross-section with concave and convex features; Step 6: Combine the warp and weft yarn fiber bundle models to construct a fabric unit cell model according to the weaving pattern of the selected fabric, and use an interference elimination algorithm to achieve interference elimination and gap insertion in the tiny areas of the model, and finally complete the reconstruction of the fabric unit cell model.

2. The method for establishing a high-fidelity unit cell model of a carbon fiber fabric according to claim 1, characterized in that: The winding rope algorithm based on the force offset principle is as follows: First, establish a closed flexible rope that wraps the set of fiber filaments, and discretize the rope into a chain of spheres composed of a series of spheres with a radius of r; virtual tensile and bending springs are set between the spheres, and their corresponding spring characteristics are assigned to maintain the smoothness and continuity of the sphere chain during movement; apply a centripetal force to the flexible rope to make the discrete small balls on it move towards the center of the area surrounded by the flexible rope at the same time; at the same time, set the repulsive force related to the overlap degree between the outer small balls and the fibers; During the movement, calculate the resultant force acting on all small balls, including the tensile spring force, bending spring force, repulsive force and centripetal force. According to the resultant force acting on them, assign a small displacement in the same direction. Finally, stop the movement when the resultant force of all small balls reaches equilibrium; when the distance between the centers of adjacent discrete spheres is less than the radius r of the sphere, perform overlapping deletion, and when the distance between adjacent centers is greater than 2r, perform sphere interpolation and supplementation. Finally, wrap the discrete fiber filaments in the same bundle. When the overall ball chain reaches the equilibrium state with the minimum potential energy, stop the movement of the spheres and output the fiber cross-section profile; perform surface interpolation on the fiber bundle profile to reconstruct the complete warp and weft fiber bundles in the fabric unit cell.

3. The method for establishing a high-fidelity unit cell model of a carbon fiber fabric according to claim 1, characterized in that: The interference elimination algorithm process is as follows: Discretize the surface of the fiber bundle into a dense point cloud set, define the in-plane space region as Ω, and at the same time define the surface of a bundle of fiber bundles as the zero potential surface Φ0. Construct the Euclidean distance field. Define the potential energy of the points inside the fiber bundle as negative, and vice versa. The magnitude of the potential energy is related to the distance d between the point p and the zero potential surface, that is: Φ = -kd, p ∈ Ω First, for the point with negative potential energy, the minimum distance from the zero potential energy surface is perpendicular to the unit normal vector corresponding to the corresponding point on the zero potential energy surface and points outward from the surface. Solve to move the point along When all the points with negative potential energy are moved to make the potential energy zero, the interference between this fiber bundle and other fiber bundles is eliminated; then, judge the potential energy of the points with non-zero potential energy. When the potential energy is greater than the set value Φ t keep the position, otherwise move the point along the normal direction corresponding to the zero potential energy surface of the fiber to which it belongs by a distance s until the potential energy is greater than the set value Φ t to achieve the insertion and control of the fiber bundle gap, with the minimum gap being d min That is: Φ t = k*d min .

4. The method for establishing a high-fidelity unit cell model of a carbon fiber fabric according to claim 2, characterized in that: The spring characteristic is the restoring force of the tensile or bending spring.

Images