Cranial-brain interface modeling method based on thick-shell unit

Abstract

The invention relates to the technical field of cranial-brain interface modeling, and discloses a cranial-brain interface modeling method based on a thick-shell element, comprising: loading a head finite element model to clean cranial-brain interface nodes, generating a thick-shell element to simulate a boundary cell layer between a dura mater and an arachnoid, and optimizing mesh quality; the thick shell unit is endowed with orthotropic material properties; setting the maximum principal stress as a failure criterion; the failure threshold is optimized through simulation-comparison-optimization circulation and machine learning by using classical corpse test data; iteratively determining an optimal failure threshold value and verifying the reliability; the technical problem that existing cranial-brain interface modeling cannot meet accurate and efficient requirements is solved.

Description

Technical Field

[0001] This invention relates to the field of cranial-brain interface modeling technology, and more specifically to a cranial-brain interface modeling method based on thick-shell units. Background Technology

[0002] Traumatic brain injury (TBI) is one of the leading causes of death and disability in traffic accidents such as vehicle collisions. To gain a deeper understanding of the mechanisms of TBI and optimize protective designs, researchers need to simulate the mechanical response of the cranio-brain structure during a collision.

[0003] Currently, the modeling methods for the brain-head interface mainly rely on the contact algorithm and the shared node method, both of which are based on traditional shell units or solid units.

[0004] The contact algorithm simulates the contact-separation behavior between two independent structures, namely the dura mater and the arachnoid mater. By defining the contact pair, friction coefficient, and contact stiffness, the contact algorithm can dynamically calculate the normal pressure and tangential shear force between the interfaces.

[0005] However, the key parameters of the contact algorithm are complex to set, requiring researchers to repeatedly debug and determine the parameters, which is time-consuming and labor-intensive. Under dynamic loads, the contact state at the brain-head interface can cause drastic fluctuations in contact force, which can easily lead to instability during the calculation process, affecting computational efficiency and accuracy.

[0006] The shared-node method simplifies computation by sharing corresponding nodes between the dura mater and arachnoid mater, eliminating relative movement between them and thus avoiding the instability of contact algorithms. This method offers high computational stability, simple setup, and is suitable for scenarios requiring rapid computation.

[0007] However, the common-node method "merges" the dura mater and arachnoid mater into a single structure, which simplifies the structural features of the cranium and makes it difficult to truly reflect the complex mechanical behavior of the dura mater and arachnoid mater, thus limiting the application scope of this technology.

[0008] The limitations of contact algorithms and shared-node methods have led to a dilemma in brain-head interface modeling where accuracy and efficiency cannot be simultaneously achieved. Summary of the Invention

[0009] The present invention aims to provide a brain-head interface modeling method based on thick-shell units to solve the technical problem that existing brain-head interface modeling cannot meet the requirements of accuracy and efficiency.

[0010] To achieve the above objectives, the present invention adopts the following technical solution: a brain-head interface modeling method based on thick-shell units, comprising: S1. Load the finite element model of the head with cranio-brain structure, determine the cranio-brain interface, which is the boundary between the dura mater and the arachnoid mater, including the geometric position and node information of the dura mater and the arachnoid mater; clean up the nodes of the cranio-brain interface. S2. Generate a thick-shell unit at the cranio-brain interface. The thick-shell unit is used to simulate the boundary cell layer between the dura mater and the arachnoid mater. The thick-shell unit is generated using the ELEMENT_TSHELL_8N keyword and the corresponding nodes of the dura mater and the arachnoid mater are connected. Check the mesh quality of the thick-shell unit and optimize it to a distortion-free mesh. S3. Assign material properties to the thick shell unit, set the material property of the thick shell unit to elastic anisotropic material, with the keyword MAT_ORTHOTROPIC_ELASTIC, and simulate the anisotropic mechanical properties of the boundary cell, including elastic modulus and shear modulus in different directions. S4. Set aging conditions for the thick shell unit. Using the LS-DYNA failure keyword MAT_ADD_EROSION, select the maximum principal stress σ of the thick shell unit as the failure criterion. When the maximum principal stress σ exceeds the failure threshold σ_fail, delete the unit to simulate material failure. S5. Using classic cadaver test data, through a cycle of simulation, comparison, and optimization, the failure threshold σ_fail of the thick shell unit is optimized and inversely calculated. S6. Based on step S5, perform iterative calculations. When the CORA score changes tend to stabilize or reach its maximum value, terminate the iteration and determine the optimal failure threshold. S7. Write the optimal failure threshold parameters into the model, perform craniocerebral impact simulation verification, compare the consistency between the simulation results and the cadaver test data, and verify the reliability of the modeling method.

[0011] The principle and advantages of this scheme are as follows: by introducing a thick shell unit (*ELEMENT_TSHELL_8N) as an intermediate layer between the dura mater and the arachnoid mater, the boundary cell layer is simulated, and the relative motion behavior between the dura mater and the arachnoid mater is indirectly described through the mechanical properties of the thick shell unit.

[0012] Thick-shell elements connect the corresponding nodes of the dura mater and arachnoid mater. Their thickness direction corresponds to the cranio-brain interface, and their planar direction corresponds to the meningeal plane. Geometrically, they adapt to extremely thin gaps, avoiding mesh distortion caused by solid elements. Anisotropic material properties are set using `MAT_ORTHOTROPIC_ELASTIC` (`MAT_002`) to simulate the tensile and shear characteristics of the boundary cell layer; the maximum principal stress failure criterion is set using `*MAT_ADD_EROSION` to simulate interface damage. As a continuous medium, the thick-shell elements avoid numerical oscillations caused by abrupt "contact-separation" state changes in contact algorithms, while preserving the structural characteristics of the cranio-brain interface and resolving the "oversimplification" problem of shared-node methods.

[0013] This approach eliminates the need to define complex parameters such as contact pairs and friction coefficients. By replacing the dynamic "contact-separation" calculation with continuous force transmission through thick-shell elements, it completely resolves the computational instability caused by contact force fluctuations. It preserves the gap characteristics and anisotropic mechanical properties of the cranio-brain interface, avoiding the mechanical behavior distortion caused by the "merging structures" of the common-node method. The thick-shell elements geometrically adapt to the extremely thin space between the meninges, avoiding mesh distortion and small time steps caused by excessively small thickness in solid elements, ensuring the overall computational efficiency and numerical stability of the model. Through the *MAT_ADD_EROSION failure keyword, the thick-shell elements can describe the interface failure process while maintaining continuous force transmission, enabling the simulation of cranio-brain interface damage behavior.

[0014] Preferably, as an improvement, the cleaning of the cranio-brain interface nodes in S1 includes: When the head model is a contact model, the number of nodes on the dura mater and arachnoid mater is extracted. If the number of nodes is inconsistent, it is adjusted to make the number of nodes on the surface of the two consistent. When the head model is modeled with common nodes, the node coordinates on the dura mater and arachnoid mater are calculated, overlapping nodes are completely separated, and the node spacing between the two is adjusted to 0.3mm-0.5mm.

[0015] The benefits of this improvement are as follows: In traditional basic modeling, a consistent number of nodes ensures that thick-shell elements can correctly connect the dura mater and arachnoid nodes, avoiding "distortion" or "invalidity" of thick-shell elements due to node mismatch, thus guaranteeing mesh quality and improving computational stability. Separating overlapping nodes and adjusting the spacing to physiological range provides a realistic geometric basis for thick-shell elements, avoiding the simplification problem of "zero gap" in the shared-node method, and preserving the structural authenticity of the craniosynostosis interface.

[0016] Preferably, as an improvement, in step S3, the basic parameters of the elastic material of the thick shell unit are consistent with the dura mater of the head model.

[0017] The beneficial effects of this improvement are: ensuring the continuity of material properties and avoiding computational oscillations caused by abrupt changes in the material properties of the dura mater and the boundary cell layer. Furthermore, since the boundary cell layer and the dura mater are physiologically continuous structures, consistent material parameters more realistically reflect the mechanical transmission process at the cranio-brain interface, balancing structural realism with computational stability.

[0018] Preferably, as an improvement, in step S3, the material parameter AOPT of the thick shell unit is adjusted to 0.0, and the material principal coordinates of the thick shell unit are set to three orthogonal principal axes, namely directions a, b, and c, where direction a is the plane extension direction, direction b is the plane transverse direction, and direction c is the thickness direction of the thick shell unit.

[0019] The beneficial effects of this improvement are: by defining orthogonal principal axes, the anisotropic mechanical properties of the boundary cell layer can be accurately simulated: the elastic moduli (E_a, E_b) in the a and b directions reflect the tensile resistance of the meningeal plane, the elastic modulus (E_c) in the c direction reflects the compressive properties of the gap; and the shear modulus (G_ab) in the ab plane reflects the shear deformation of the meningeal plane. This setting ensures that the mechanical response of the thick-shell unit is consistent with the physiological structure, thus improving the realism of the model.

[0020] Preferably, as an improvement, in step S3, the material density ρ of the thick-shell unit is further set to 1010. The elastic modulus in the three orthogonal principal axis directions is E_a=E_b=E_c=0.0315GPa, and the Poisson's ratio PR_ab in the ab plane is 0.35.

[0021] The beneficial effects of this improvement are: the density parameter is consistent with the density of brain tissue, ensuring the accuracy of inertial force calculation; the elastic modulus of 0.0315 GPa is close to the physiological modulus of the boundary cell layer, ensuring the realism of tensile / compressive deformation; and the Poisson's ratio PR_ab = 0.35 reflects the lateral contractile characteristics of the meningeal plane, accurately simulating the shear behavior of the ab plane when combined with the shear modulus calculation formula. Overall, this parameter setting achieves a balance between simplified calculation and realistic mechanical properties, avoiding complex parameter adjustments and improving computational efficiency and stability.

[0022] Preferably, as an improvement, in step S5, by setting the initial failure threshold to 90 kPa, setting the simulation matrix, generating multiple sets of simulations with different failure thresholds, calculating the CORA score values ​​of each simulation result curve and the experimental curve, and using the maximization of the CORA score value as the objective function, a machine learning-driven optimization algorithm is used to establish a surrogate model for solving.

[0023] The beneficial effects of this improvement are: the over-simulation matrix covers the threshold range, replacing the traditional "trial and error method" and reducing manual intervention. The CORA score quantifies the consistency between the simulation and experimental curves; the machine learning surrogate model quickly fits the "threshold-CORA" relationship to find the optimal threshold, i.e., maximizing CORA's σ_fail, ensuring the reliability of damage simulation. The surrogate model replaces a large number of repetitive simulations, shortening the optimization cycle and solving the problem of cumbersome parameter debugging in contact algorithms.

[0024] Preferably, as an improvement, the classic cadaver test data used in S5 is the classic intracranial injury test Nahum et al. 1977.

[0025] The beneficial effects of this improvement are: the data is classic and authoritative data in traumatic brain injury research, with widespread acceptance; the failure threshold optimized based on this data is more universal; the consistency between simulation results and experiments verifies the reliability of the model, ensuring the accuracy and persuasiveness of the solution in practical applications. Attached Figure Description

[0026] Figure 1 This is a flowchart of an embodiment of the present invention.

[0027] Figure 2 This is a schematic diagram of the coordinate axis of the thick shell unit in an embodiment of the present invention. Detailed Implementation

[0028] The following detailed description illustrates the specific implementation method: Example The basics are as follows: Figure 1 As shown, a brain-head interface modeling method based on thick-shell units includes: S1. Load a finite element model of the head with cranial-brain structure, such as in LS-DYNA format, and determine the cranial-brain interface from the head finite element model. The cranial-brain interface is between the two layers of meninges, the dura mater and the arachnoid mater, and includes the geometric location, ID, and nodes of the dura mater and the arachnoid mater.

[0029] Based on the modeling type of the cranio-brain interface head model, the nodes of the cranio-brain interface are cleaned up. When the cranio-brain interface is a contact-modeled head model, the number of nodes on the dura mater and arachnoid mater is extracted, and it is determined whether the number of nodes on the dura mater and arachnoid mater is consistent. If they are inconsistent, adjustments are made to make the number of nodes on the surfaces of the dura mater and arachnoid mater consistent. Adjustment methods include inserting virtual nodes or deleting redundant nodes.

[0030] When the head model with shared nodes is used for the craniotomy interface, the coordinates of nodes on the dura mater and arachnoid mater are calculated. Overlapping nodes on the dura mater and arachnoid mater are completely separated, and the node spacing between the dura mater and arachnoid mater is adjusted to 0.3mm-0.5mm. Spacing adjustment methods include coordinate translation.

[0031] S2. Generate thick-shell elements at the craniosynaptic interface. These elements simulate the boundary cell layer between the dura mater and the arachnoid mater. Use the *ELEMENT_TSHELL_8N keyword to generate thick-shell elements and connect the corresponding nodes of the dura mater and the arachnoid mater. Check the mesh quality and optimize to a distortion-free mesh.

[0032] Mesh quality quantification metrics include aspect ratio, Jacobian determinant, distortion rate, warpage, and node correspondence. Thick-shell elements can be selected using HyperMesh to generate a quality report and visualize the location of problem areas.

[0033] When aspect ratio / Jacobi issues exist, for "slender" or "twisted" thick-shell units, adjust the corresponding node positions of the dura mater / arachnoid membrane to make the unit closer to a "regular quadrilateral".

[0034] When there are node mismatches or large-area distortion issues, if a region's thick shell elements are distorted due to dura mater / arachnoid node mismatches, the mesh for that region needs to be re-generated. Delete the thick shell elements in the problematic region and their corresponding dura mater / arachnoid shell elements. Use the "Mesh→2D→Automesh" tool to re-mesh the dura mater / arachnoid shell elements with a more uniform node density. Regenerate the thick shell elements, ensuring that the upper and lower nodes correspond one-to-one.

[0035] S3. Assign material properties to the thick-shell unit; since the boundary cell material exhibits anisotropy, set the material properties of the thick-shell unit to elastic anisotropic material, with the keyword *MAT_002 (*MAT_ORTHOTROPIC_ELASTIC); simulate the mechanical properties of the boundary cell, such as elastic modulus and shear modulus in different directions.

[0036] Adjust the material parameter AOPT=0.0, nodes 1 to 4 are arachnoid nodes, nodes 5 to 8 are dura mater nodes, directions a, b, and c are the three orthogonal principal axes of the material, and direction c is the thickness direction of the thick shell element.

[0037] Assign orthotropic elastic material properties, including density ρ, elastic modulus E_i, Poisson's ratio PR_ij, and shear modulus G_ij.

[0038] Density ρ is the mass per unit volume, used in dynamic analysis to calculate inertial forces.

[0039] The elastic modulus E_i is the tensile / compressive elastic modulus of a material in the i-direction, i.e., the direction of the material's principal axis. The unit is MPa or GPa. It describes the material's ability to resist elastic deformation in the i-direction. The larger E_i is, the harder the material is.

[0040] Poisson's ratio PR_ij refers to the ratio of the transverse strain in the j-direction to the axial strain in the i-direction when a material is stretched in the i-direction. A larger PR_ij indicates that the shrinkage in the j-direction is more pronounced when stretched in the i-direction.

[0041] The shear modulus G_ij refers to the material's ability to resist shear deformation within the ij plane, and its unit is MPa or GPa. The larger the G_ij, the more difficult the material is to be sheared.

[0042] According to the generalized Hooke's law in elasticity, the shear modulus can be calculated from the elastic modulus and Poisson's ratio: ; Set the basic elastic material parameters of the thick shell element to be consistent with the dura mater of the model; set PR_ab to be large and PR_bc=PR_ac to be small.

[0043] Appendix Figure 2 In the diagram, 'a' and 'b' are perpendicular to each other. 'a' represents the "planar extension direction" of the thick-shell unit, such as along the anterior-posterior direction of the meninges. The larger E_a is, the more difficult it is for the thick-shell unit to be stretched in the 'a' direction. 'b' represents the "planar transverse direction" of the thick-shell unit, such as along the lateral direction of the meninges. 'c' represents the "thickness direction" of the thick-shell unit, such as perpendicular to the meningeal plane, corresponding to the gap direction at the cranio-brain interface. 'v_12' represents the vectors from nodes 1 to 2, and 'v_14' represents the vectors from nodes 1 to 4.

[0044] In this embodiment, the density ρ is set to 1010. The elastic modulus in the three orthogonal principal axis directions is E_a=E_b=E_c=0.0315GPa, and the Poisson's ratio PR_ab in the ab plane is 0.35.

[0045] Based on the above formula for calculating shear modulus, the shear modulus of the ab plane is approximately 11.67 MPa.

[0046] S4. Set aging conditions for the thick shell element. Using the LS-DYNA failure keyword *MAT_ADD_EROSION, select the maximum principal stress σ of the element as the failure criterion. When the maximum principal stress σ exceeds the failure threshold σ_fail, the element is deleted to simulate material failure.

[0047] S5. Using classic cadaver test data, such as the classic intracranial injury test Nahum et al. 1977, the failure threshold of the thick shell element is optimized and reverse-calculated through a "simulation-comparison-optimization" cycle.

[0048] Using the maximum principal stress failure threshold σ_fail as the design variable, the initial failure threshold was set to 90 kPa. A simulation matrix was set, and multiple simulations with different σ_fail values ​​were generated to cover the possible threshold range.

[0049] The *CORRELATE command is used to calculate the CORA scores of each simulation result curve and the experimental curve. Maximizing the CORA score f(σ_fail) is used as the objective function, and a machine learning-driven optimization algorithm is employed to establish a surrogate model for solving the problem. The optimization problem is expressed as follows: ; One approach is to use a surrogate model, such as Kriging or a neural network, to fit the relationship between "σ_fail and CORA" and quickly find the maximum σ_fail in CORA, thereby improving simulation efficiency.

[0050] S6. Repeat the iteration. Based on step S5, perform iterative calculations. The iteration is terminated when the CORA score changes tend to stabilize or reach its maximum value, thus determining the optimal failure threshold for the thick-shell element.

[0051] S7. Model Verification. The obtained thick-shell element failure threshold parameters are written into the model, and a craniocerebral impact simulation is performed for verification. The consistency between the simulation results and the response of the cadaver test is compared to verify the reliability and applicability of the thick-shell element craniocerebral interface modeling method.

[0052] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for modeling the brain-head interface based on thick-shell units, characterized in that, include: S1. Load the finite element model of the head with cranio-brain structure, determine the cranio-brain interface, which is the boundary between the dura mater and the arachnoid mater, including the geometric position and node information of the dura mater and the arachnoid mater; clean up the nodes of the cranio-brain interface. S2. Generate a thick-shell unit at the cranio-brain interface. The thick-shell unit is used to simulate the boundary cell layer between the dura mater and the arachnoid mater. The thick-shell unit is generated using the ELEMENT_TSHELL_8N keyword and the corresponding nodes of the dura mater and the arachnoid mater are connected. Check the mesh quality of the thick-shell unit and optimize it to a distortion-free mesh. S3. Assign material properties to the thick shell unit, set the material property of the thick shell unit to elastic anisotropic material, with the keyword MAT_ORTHOTROPIC_ELASTIC, and simulate the anisotropic mechanical properties of the boundary cell, including elastic modulus and shear modulus in different directions. S4. Set aging conditions for the thick shell unit. Using the LS-DYNA failure keyword MAT_ADD_EROSION, select the maximum principal stress σ of the thick shell unit as the failure criterion. When the maximum principal stress σ exceeds the failure threshold σ_fail, delete the unit to simulate material failure. S5. Using classic cadaver test data, through a cycle of simulation, comparison, and optimization, the failure threshold σ_fail of the thick shell unit is optimized and inversely calculated. S6. Based on step S5, perform iterative calculations. When the CORA score changes tend to stabilize or reach its maximum value, terminate the iteration and determine the optimal failure threshold. S7. Write the optimal failure threshold parameters into the model, perform craniocerebral impact simulation verification, compare the consistency between the simulation results and the cadaver test data, and verify the reliability of the modeling method.

2. The brain-head interface modeling method based on thick-shell units according to claim 1, characterized in that, The cleaning of the cranio-brain interface nodes in S1 includes: When the head model is a contact model, the number of nodes on the dura mater and arachnoid mater is extracted. If the number of nodes is inconsistent, it is adjusted to make the number of nodes on the surface of the two consistent. When the head model is modeled with common nodes, the node coordinates on the dura mater and arachnoid mater are calculated, overlapping nodes are completely separated, and the node spacing between the two is adjusted to 0.3mm-0.5mm.

3. The brain-head interface modeling method based on thick-shell units according to claim 2, characterized in that: In S3, the basic parameters of the elastic material of the thick shell unit are consistent with those of the dura mater of the head model.

4. The brain-head interface modeling method based on thick-shell units according to claim 3, characterized in that: In step S3, the material parameter AOPT of the thick shell element is adjusted to 0.0, and the material principal coordinates of the thick shell element are set to three orthogonal principal axes, namely directions a, b, and c. Direction a is the plane extension direction, direction b is the plane transverse direction, and direction c is the thickness direction of the thick shell element.

5. The brain-head interface modeling method based on thick-shell units according to claim 4, characterized in that: In S3, the material density ρ of the thick shell unit is also set to 1010. The elastic modulus in the three orthogonal principal axis directions is E_a=E_b=E_c=0.0315GPa, and the Poisson's ratio PR_ab in the ab plane is 0.

35.

6. The brain-head interface modeling method based on thick-shell units according to claim 5, characterized in that: In step S5, by setting the initial failure threshold to 90 kPa, setting the simulation matrix, generating multiple sets of simulations with different failure thresholds, calculating the CORA score values ​​of each simulation result curve and the experimental curve, and using the maximization of the CORA score value as the objective function, a machine learning-driven optimization algorithm is used to establish a surrogate model for solving.

7. A brain-head interface modeling method based on thick-shell units according to claim 6, characterized in that: The classic cadaveric test data used in S5 are from the classic intracranial injury test Nahum et al. 1977.

Images