An equal-wall-thickness gastroesophageal reverse modeling and convergent fluid-structure coupling simulation method

Abstract

This invention discloses a method for inverse modeling and convergent fluid-structure interaction (FSI) simulation of the gastroesophageal tract with uniform wall thickness, belonging to the interdisciplinary field of medical 3D modeling and biomechanical simulation. This method uses multi-slice CT images as the data source, extracts organ contour control points from dual perspectives, and constructs a global, uniform-wall-thickness dual-solid model of the gastroesophageal tract through lofting reconstruction, curvature adaptive smoothing, and bidirectional isometric shelling. After topology cleaning and end-to-end connection, a standardized cavity solid is obtained. After modeling, equivalent displacement boundaries are used to characterize the three physiological narrowings of the esophagus, combined with Yeoh hyperelastic constitutive models to describe soft tissue deformation. An innovative fluid transient and solid steady-state unidirectional partitioned FSI strategy is adopted, effectively solving the defects of traditional global coupling such as easy divergence, uneven wall thickness in large curvature models, and mesh distortion in narrowing areas. This invention significantly improves the anatomical geometric fidelity and biomechanical simulation stability of the gastroesophageal tract, and is suitable for the mechanical mechanism analysis of normal digestive tract, reflux esophagitis, Barrett's esophagus, and hiatal hernia, providing a reliable digital simulation tool for biomechanical research in the precise diagnosis and treatment of the digestive tract.

Description

Technical Field

[0001] This invention belongs to the field of medical 3D modeling and biomechanical finite element simulation technology, specifically involving a reverse modeling and convergent fluid-structure interaction simulation method for the gastroesophageal tract with equal wall thickness, which is applicable to the digital reconstruction of digestive organ anatomy and the simulation analysis of the mechanical mechanism of reflux disease. Background Technology

[0002] Gastroesophageal reflux disease (GERD) is a common chronic disease in gastroenterology. The persistent reflux of gastric acid and bile irritates the esophageal mucosa, gradually leading to inflammation and erosion. Prolonged GERD can progress to Barrett's esophagus, carrying a risk of esophageal adenocarcinoma. Clinically, non-erosive reflux disease accounts for 60-70% of all cases, presenting only with subjective symptoms of acid reflux and burning pain, without organic mucosal damage. Reflux esophagitis accounts for 20-30%, accompanied by mucosal congestion, erosion, and even superficial ulcers. Barrett's esophagus accounts for 5-10%, a precancerous metaplastic lesion where squamous epithelium is replaced by columnar epithelium. The three types of lesions show significant differences in mucosal mechanical tolerance thresholds and reflux load response characteristics. Clinical endoscopic observation alone cannot quantify the distribution of internal esophageal stress and shear force; therefore, high-precision digital biomechanical simulation models are necessary for mechanistic research.

[0003] Current digital modeling and simulation technologies for gastroesophageal biomechanical analysis suffer from two major bottlenecks: geometric modeling defects and numerical solution stability defects. First, at the geometric modeling level: traditional modeling often uses only single-view CT slices to extract organ contours. High-curvature anatomical structures such as the greater curvature of the stomach, pylorus, and esophageal physiological narrowings have blurred boundaries in single-slice imaging, resulting in significant geometric deficiencies in the extracted contours. Conventional unidirectional shelling operations only offset the surface along a single-sided normal, leading to sudden increases or decreases in local wall thickness in areas of severe curvature, and in severe cases, surface penetration, failing to replicate the uniform physiological structure of the human digestive tract wall. Some modeling schemes directly stretch and cut solid geometry to simulate the three physiological narrowings: the cricopharyngeal muscle, the aortic crossing, and the diaphragmatic hiatus. Geometric modifications cause abrupt changes in mesh size and severe element distortion in the narrowing areas, significantly reducing simulation accuracy. Second, at the numerical simulation level: the soft tissue of the gastroesophageal tract has superelastic, large deformation, and near-incompressible nonlinear mechanical properties. The flow of gastric juice is an unsteady transient fluid. The traditional global unified fluid-structure interaction scheme solves the fluid and solid fields simultaneously and iteratively. The superposition of the time-varying fluid load and the large deformation of the soft tissue with the dual nonlinear effects easily causes iterative oscillation and calculation divergence. Even after multiple parameter adjustments, it is still difficult to converge, resulting in extremely low computational efficiency and data reliability.

[0004] In existing publicly available technologies, standalone CT reverse modeling schemes cannot simultaneously ensure the continuity of large-curvature surfaces and the uniform wall thickness characteristics across the entire domain; conventional fluid-structure interaction (FSI) simulation algorithms lack optimized solution frameworks for digestive tract soft tissues. Therefore, these two approaches cannot form a complete and complementary technical system, making it difficult to simultaneously meet the dual requirements of anatomical geometric fidelity and numerical computational stability. Thus, there is an urgent need to develop an integrated method for gastroesophageal reverse modeling and highly convergent FSI simulation to eliminate modeling errors at the source of geometric reconstruction and to solve simulation divergence problems through solution framework optimization. Summary of the Invention

[0005] To address the multiple technical shortcomings of existing digital gastroesophageal modeling methods, such as loss of large-curvature anatomical features during single-view contour extraction, uneven wall thickness during unidirectional shelling, mesh distortion caused by geometric modification leading to stenosis, and the tendency for iterative divergence in traditional global fluid-structure interaction (FSI) solutions, this invention provides a method for inverse modeling of the gastroesophageal tract with uniform wall thickness and convergent FSI simulation. This invention establishes a complete technical chain of "CT dual-view contour extraction - multi-section lofting reconstruction - curvature smoothing bidirectional equidistant shelling - topological lumen cleaning - partitioned unidirectional coupling simulation," eliminating the need for dedicated high-precision scanning equipment and enabling high-fidelity digestive tract solid modeling solely based on clinically available CT images. It innovatively employs equivalent displacement boundaries to simulate physiological stenosis and a fluid transient / solid steady-state partitioned unidirectional coupling strategy, simultaneously improving the accuracy of the geometric model and the convergence of simulation calculations, making it suitable for biomechanical quantitative analysis of various normal and pathological gastroesophageal samples.

[0006] This invention is divided into two main technical modules: high-precision inverse geometric modeling with uniform wall thickness and partitioned convergent fluid-structure interaction simulation. The innovative mechanisms of each module are described below: 1. Innovative Principles of the Geometric Modeling Module (1) Dual-view contour complementary extraction: a single coronal section can only capture the complete contour of the stomach, while an axial section clearly presents the longitudinal morphology of the esophagus. The two types of sections are combined to extract control points, which complement each other to compensate for the lack of boundary information of high curvature in a single view and completely restore the anatomical shape of the entire gastroesophageal segment. (2) Curvature adaptive G2 smoothing + bidirectional synchronous shelling: Traditional unidirectional shelling only offsets the surface on one side, and the curvature change area will produce geometric distortion. This invention generates equidistant surfaces synchronously along the inner and outer normals of the model, and maintains a uniform pipe wall thickness throughout the entire section. Combined with curvature and deflection dual constraint smoothing, the modeling noise is eliminated while retaining the three key geometrical points of physiological stenosis, and the surface smoothness reaches the finite element simulation standard. (3) Equivalent narrow representation of displacement boundary: abandoning the traditional narrow modeling method of cutting and stretching solids, the swallowing cycle deformation is replicated by applying a displacement field that matches the tissue compression effect to the corresponding anatomical region without changing the original cavity geometry, thus avoiding the mesh distortion problem in the narrow region from the root.

[0007] 2. Innovative Principles of Fluid-Structure Coupling Simulation Module The gastroesophageal system comprises two physical fields: transient gastric fluid and hyperelastic, large-deformation esophageal soft tissue. Simultaneous iteration across the entire domain results in superimposed double nonlinearities, leading to numerical instability. This invention separates the solution sequence for the two fields: a separate transient solution for the fluid domain to capture time-varying pressure and shear stress distributions; and a separate steady-state solution for the solid domain to obtain the equilibrium stress and strain of the esophagus under fluid loads. Only the fluid calculation results are unidirectionally mapped to the solid mesh, eliminating bidirectional iterative interaction, significantly weakening the oscillations caused by nonlinear coupling, and significantly improving the iteration convergence probability. Simultaneously, the Yeoh hyperelastic strain potential energy model is used to accurately match the near-incompressible, large-deformation mechanical characteristics of the soft tissue, ensuring the physical realism of the simulation.

[0008] Specifically, the technical solution adopted in this invention is: a method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with equal wall thickness, which includes the following core steps: Step 1: Joint extraction of dual-view CT tomographic contours Healthy volunteers and subjects with reflux esophagitis or Barrett's esophagus underwent multi-slice CT scans with a uniform slice interval of 1.25 mm. Two sets of image sequences were acquired simultaneously: coronal and axial. Using the ClosedCurve tool in medical image processing software, the complete outer contour of the stomach was delineated on the coronal slices and the complete outer contour of the esophagus was delineated on the axial slices, and two sets of discrete coordinate control points were exported. The discrete points were fitted with cubic B-spline curves to smooth the continuous organ boundaries. After unifying the two sets of contour space coordinate systems, they were merged and archived to construct a complete dataset of gastroesophageal anatomical control points.

[0009] The mathematical expression for the cubic B-spline curve is: In the formula: N i,3 (u) Let P be a cubic B-spline basis function. i The coordinates of the discrete control points for each cross-sectional profile are given.

[0010] Step 2, Initial 3D Solid Reconstruction via Multi-section Lofting All contour control points after unifying coordinates are imported into the 3D CAD modeling software in batches. Based on the spatial topological relationship of each cross-sectional contour, the Loft surface fitting algorithm is adopted to stitch the solid layer by layer with layered spline curves as guides. The lofting process ensures that there are no surface abrupt changes or geometric breaks in the high curvature transition areas such as the greater curvature of the stomach, pylorus, and esophageal stenosis, generating a continuous and smooth initial 3D solid model of the stomach and esophagus.

[0011] Step 3: Composite modeling of curvature adaptive smoothing and bidirectional equidistant shelling. The initial solid generated by lofting is imported into reverse engineering software for surface optimization and two-layer solid construction, fully encompassing three layers of processing logic: 3.1 Curvature Adaptive Global G2 Smoothing Processing: Feature removal, random noise removal, mesh relaxation, and spike removal operations are performed on all surfaces of the solid. By constraining the continuity of the surface through the curvature and torsion index of the spatial curve, a global G2 level smooth transition is achieved while preserving the key geometric features of the three physiological stenosis points: cricopharyngeal muscle, aortic crossing, and diaphragmatic hiatus. Formula for calculating the curvature of a space curve: Formula for calculating the deflection of a space curve: In the formula: C(t) represents the space curve vector function.

[0012] 3.2 Bidirectional Synchronous Isotropic Shelling: Based on the measured thickness of the gastric and esophageal wall in the human body, the offset distance is set, and equidistant offset surfaces are synchronously generated along the unit normal vectors of the inner and outer surfaces of the solid, constructing a two-layer nested composite model of the outer wall solid and the inner wall surface; General expression for the bidirectional equidistant surface: In the formula: S(u,v) The original surface parametric equations, S offset (u,v) For the offset surface, d This represents the offset thickness of the pipe wall. n(u,v) is the unit normal vector of the surface.

[0013] 3.3 Anatomical partitioning: Anatomical boundaries are defined for the outer solid and inner curved surfaces respectively; the outer layer is given a dedicated contour line for three physiological stenosis, and the inner layer is divided into partitions according to the natural direction of the mucosa, providing a geometric reference for the subsequent precise loading of finite element mechanical boundaries.

[0014] Step 4, Topology cleaning and end-of-tube clearance treatment The two-layer composite model was re-imported into 3D CAD software for topology optimization and cavity forming, which was done in two stages: 4.1 Redundant Topology Removal: Boolean operations are used to remove auxiliary surfaces and overlapping patches that have no physiological significance generated during the modeling process, eliminating non-manifold topological conflicts caused by multi-source geometric fusion. The model ultimately retains only a single, closed, watertight solid base. 4.2 Directional resection of the lumen: Based on anatomical landmarks, locate the cricopharyngeal muscle at the upper end of the esophagus and the pyloric region at the distal end, perform directional solid resection, and reconnect the entire digestive tract lumen to restore the physiological structure of the continuous gastroesophageal passage. Divide the geometric boundaries of the independent fluid computational domain and the solid structure domain to meet the geometric requirements before finite element mesh generation.

[0015] Step 5: Partitioned unidirectional fluid-structure interaction finite element simulation modeling A biomechanical simulation model is built based on the cleaned, uniformly thick cavity solid, including three core operations: boundary setting, material definition, and partitioned solution. 5.1 Physiological Stenosis Equivalent Displacement Boundary Loading: Three physiological stenosis areas in the esophagus are precisely located, and a displacement field matching the periodic compression effect of the surrounding soft tissue is applied to dynamically replicate the deformation and stress concentration characteristics of the stenosis areas during swallowing; Nonlinear displacement deformation expression: In the formula: I is the unit tensor. The displacement gradient tensor in the reference coordinate system; 5.2 Constitutive Definition of Soft Tissue Materials: The Yeoh hyperelastic strain potential energy model was used to characterize the large deformation and near-incompressible nonlinear mechanical properties of gastroesophageal soft tissue. Constitutive parameters of the material were fitted through in vitro tensile tests. Yeoh strain potential energy expression: Second Piola-Kirchhoff stress in soft tissue: In the formula: W is the strain potential energy, I1 is the first strain invariant, and C i0 Here, I represents the material fitting coefficients, and I represents the unit tensor. 5.3 Partitioned Unidirectional Coupled Solution Framework: The system divides the system into independent fluid and solid domains. The fluid domain is set to a transient solution mode to solve the conservation equations for gastric fluid mass, momentum, and energy, capturing time-varying pressure and shear stress. The solid domain is set to a steady-state solution mode to receive unidirectional mapped loads from the fluid and solve for the equilibrium stress field. The fluid-to-solid area conversion, load mapping, and solid momentum balance are used to complete the full-field numerical calculation using tensor equations. The excitation is transmitted only unidirectionally, without bidirectional iterative interaction, thus suppressing computational oscillations.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The dual-view contour extraction mechanism solves the problem of loss of anatomical boundaries in high curvature single slices, greatly improves the geometric fidelity of the three-dimensional reconstruction model, and completely restores the complex physiological shape of the gastroesophageal tract. The bidirectional synchronous equidistant shelling process, unlike the traditional unidirectional unilateral offset method, maintains uniform tube wall thickness even in areas with severe curvature such as the greater curvature of the stomach and the pylorus, completely eliminating defects such as surface mold penetration and wall thickness distortion. The displacement boundary is equivalent to physiological narrowing, without modifying the solid geometry, thus avoiding mesh element distortion and deformation in narrow regions from the source, and improving the quality of finite element mesh and simulation accuracy. The partitioned unidirectional fluid-structure interaction solution framework separates the transient fluid and steady-state solid iteration logic, eliminates the bidirectional coupling interaction, weakens the numerical oscillation caused by the superposition of the dual nonlinearity of fluid and soft tissue, significantly reduces the probability of iteration divergence, and simultaneously improves computational stability and solution efficiency. The modeling data source uses only routine clinical CT images, without requiring specialized equipment such as high-precision industrial scanning turntables, making the hardware threshold low; the entire modeling and simulation process is compatible with various samples such as healthy volunteers, reflux esophagitis, Barrett's esophagus, and hiatal hernia, making it highly universal; The entire technology is integrated and seamless, directly outputting a convergent simulation model that can be used for quantitative mechanical analysis from the original medical images, eliminating the need for repeated geometric corrections using multiple software programs and significantly shortening the preprocessing cycle for digestive tract biomechanical simulation. Attached Figure Description

[0017] Figure 1 This is a general flowchart of the overall modeling and fluid-structure interaction simulation process of this invention; Figure 2 This is a schematic diagram of the contour localization of the gastroesophageal CT images of the present invention; Figure 2 (a) Schematic diagram of gastric contour extraction from coronal CT scan. Figure 2 (b) Schematic diagram of axial esophageal CT contour extraction; Figure 3 This is a diagram of the initial three-dimensional solid model of the gastroesophageal tract generated by multi-section lofting in this invention; Figure 4 This is a model diagram of the curved smooth + bidirectional equidistant shelling double-layer equal wall thickness mesh of the present invention; Figure 5 This is a solid model diagram of the hollow watertight cavity after topology clearing and lumen connection according to the present invention; Figure 6 This is a cloud map showing the three physiologically narrow loading areas and corresponding deformations of the present invention; Figure 6 (a) Schematic diagram of the loading positions of three physiological strictures in the esophagus. Figure 6 (b) Fluid-structure interaction stress-deformation contour map. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the following embodiments are only for illustrating the present invention and are not intended to limit the scope of the present invention.

[0019] Example 1 This embodiment uses conventional clinical CT images for modeling and general-purpose finite element software for fluid-structure interaction simulation. The software used is only an example, and the method of this invention is not limited to specific commercial software. This embodiment can use CT images of healthy volunteers, subjects with reflux esophagitis, or Barrett's esophagus as the data source, with a uniform slice spacing of 1.25 mm. The overall implementation process is shown in the attached figure. Figure 1 As shown A method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness includes the following steps: Step 1: Joint extraction of dual-view CT tomographic contours The image acquisition and contour extraction results in this embodiment are as follows: Figure 2 (a) Figure 2 As shown in (b), Figure 2 (a) is a schematic diagram of the coronal plane gastric CT contour extraction. Figure 2 (b) is a schematic diagram of axial esophageal CT contour extraction. Multi-slice CT images were read using medical image processing software. The outer contour of the stomach was delineated in coronal slices and the outer contour of the esophagus was delineated in axial slices, and discrete coordinate points were exported. The discrete control points were fitted with cubic B-spline curves and stitched together to obtain a complete organ contour dataset.

[0020] The mathematical expression for a cubic B-spline curve is: In the formula: N i,3 (u) Let P be a cubic B-spline basis function. i The coordinates of the discrete control points for each cross-sectional profile are given.

[0021] Step 2, Initial 3D Solid Reconstruction via Multi-section Lofting Import all calibrated contour control points into SolidWorks 3D CAD software in batches, call the Loft loft feature tool, and generate a continuous solid with the layered spline contour as the guide line; the software automatically interpolates the transition area surface of each section to eliminate the geometric discontinuity caused by contour layering, and generates a complete initial 3D solid model of the gastroesophageal tract, which fully preserves all anatomical features such as the fundus of the stomach, the greater curvature of the stomach, the pylorus, and the three narrow points of the esophagus.

[0022] Step 3: Composite modeling of curvature adaptive smoothing and bidirectional equidistant shelling. Import the SolidWorks output solid into Geomagic Wrap reverse engineering software to complete surface post-processing: 3.1 Curvature Adaptive Global G2 Smoothing Processing: Performs feature removal, random noise filtering, mesh relaxation, and spike removal; the software's built-in curvature and torsion solving module calculates the surface continuity index piece by piece, optimizing the surface to G2 level smoothness without removing three physiologically narrow geometric features; Formula for calculating the curvature of a space curve: Formula for calculating the deflection of a space curve: In the formula: C(t) represents the space curve vector function.

[0023] 3.2 Bidirectional Synchronous Equidistant Shelling: Input the measured average wall thickness offset parameters of the human stomach and esophagus. d Enable the bidirectional synchronous Shell extraction function to synchronously generate equidistant offset surfaces along the unit normal vectors of the inner and outer surfaces of the solid, forming a two-layer nested composite model of outer wall solid and inner wall surface; the general expression for the bidirectional equidistant surface is: In the formula: S(u,v) The original surface parametric equations, S offset (u,v) For the offset surface, d This represents the offset thickness of the pipe wall. n(u,v) is the unit normal vector of the surface.

[0024] 3.3 Anatomical partitioning: The outer solid layer was manually marked with the exclusive contour boundaries of three narrow points: the cricopharyngeal muscle, the aortic crossing, and the diaphragmatic hiatus. The inner curved surface was divided into independent partitions according to the natural orientation of the digestive tract mucosa, providing geometric markers for subsequent precise finite element boundary positioning.

[0025] Step 4, Topology cleaning and end-of-tube clearance treatment Re-import the two-layer composite model output by Geomagic Wrap into SolidWorks for topology optimization: 4.1 Redundant Topology Removal: Boolean difference and Boolean merge operations are used to remove auxiliary patches and overlapping surfaces generated during modeling, eliminating non-manifold defects caused by geometric transformations in multiple software programs. The final model retains only a single closed watertight entity. 4.2 Directional connection of the lumen: Based on anatomical landmarks, locate the closed solid area at the upper end of the esophagus and the distal end of the pylorus, create a resection feature to connect the entire digestive tract lumen, form a continuous hollow cavity model, distinguish the internal fluid domain and the external esophageal solid domain, and complete the preparation of the finite element geometric basis.

[0026] Step 5: Partitioned unidirectional fluid-structure interaction finite element simulation modeling Import the cavity solid into Comsol multiphysics finite element simulation software to build a biomechanical model: 5.1 Physiological Stenosis Equivalent Displacement Boundary Loading: Three physiological stenosis areas in the esophagus were precisely selected, and a displacement field matching the compression effect of surrounding tissues was applied to replicate the dynamic deformation during the swallowing cycle; Nonlinear displacement deformation expression: In the formula: I is the unit tensor. The displacement gradient tensor in the reference coordinate system; 5.2 Constitutive Definition of Soft Tissue Materials: The Yeoh hyperelastic model is used to describe the mechanical behavior of esophageal soft tissue, and the C0 model is fitted based on in vitro tensile tests. 10 C 20 C 30 Material coefficients; Yeoh strain potential energy expression: Second Piola-Kirchhoff stress in soft tissue: In the formula: W is the strain potential energy, I1 is the first strain invariant, and C i0 Here, I represents the material fitting coefficients, and I represents the unit tensor. 5.3 Partitioned Unidirectional Coupled Solution Framework: The model is divided into two major computational domains: a gastric fluid domain and an esophageal solid domain. The fluid domain is set to a transient solution mode to solve the mass conservation equation, momentum conservation equation, and energy conservation equation, capturing the time-varying pressure and shear stress field of the gastric fluid. The solid domain is set to a steady-state solution mode to receive the unidirectional mapped load at the fluid interface and solve the equilibrium stress and strain. A complete set of tensor equations, including deformation gradient, Jacobian determinant, fluid-solid area conversion, and equivalent fluid body force, are used to complete the numerical solution. The entire process involves only unidirectional excitation from the fluid to the solid, without bidirectional iterative interaction, which greatly reduces the risk of numerical oscillation and divergence.

[0027] Fluid mass conservation equation: Where ρ is the fluid density, t is time, u is the velocity vector, and ∇ is the divergence operator. Fluid momentum conservation equation: Where ρ is the fluid density, p is the static pressure, τ is the viscous stress tensor, μ is the dynamic viscosity, f is the volume force per unit mass, and I is the unit tensor.

[0028] Fluid energy conservation equation: The complete set of tensor governing equations, including the reference configuration solid momentum equation, fluid-solid area conversion formula, fluid load mapping vector, and hyperelastic solid equilibrium equation, are fully embedded in the solver to achieve stable and convergent unidirectional fluid-solid coupling numerical calculations.

[0029] In this embodiment, MicroDicom, SolidWorks, Geomagic Wrap, and Comsol are only commercial software used for practical examples. The method and process of this invention are not limited to specific software. Image, modeling, and finite element software with equivalent functions can be used to replace and implement all the steps of this invention. Parameters such as CT slice spacing, tube wall offset thickness, Yeoh material coefficient, and fluid solution time step can be flexibly adjusted according to the subject's body size and lesion type. All parameter adjustment schemes fall within the protection scope of this invention. In addition to healthy volunteers, the data source for the example can be replaced with CT images of patients with reflux esophagitis, Barrett's esophagus, or hiatal hernia. Only the contour extraction range and displacement boundary amplitude need to be finely adjusted according to the anatomical characteristics of the lesion, while the core process of the method remains completely unchanged.

[0030] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness, characterized in that, The complete execution steps are as follows: Step 1: Acquire multi-slice CT images of the human gastroesophageal tract, and simultaneously select coronal and axial slices to extract discrete control points of the organ's outer contour. Use cubic B-spline curves to fit the discrete points to generate a continuous contour. Rely on the complementary geometric information of dual-view images to compensate for the loss of features in large curvature areas of a single slice, and provide a complete anatomical control point dataset for three-dimensional entity reconstruction. Step 2: Import all contour point sets after coordinate unification into the 3D computer-aided design software. Based on the multi-section spatial topology relationship, the lofted surface fitting technology is used to stitch together the contours of each section layer by layer to generate a continuous and abrupt initial 3D solid model of the gastroesophageal tract. Step 3: Import the initial solid into the reverse engineering software to perform surface optimization. Use the curvature adaptive smoothing algorithm to eliminate modeling noise and achieve full-domain G2 surface continuity. Based on the measured thickness of the human digestive tract wall, perform bidirectional equidistant shelling operation simultaneously along the normal of the inner and outer surfaces of the model to generate a double-layer nested composite model of outer wall solid and inner wall surface, achieving uniform wall thickness throughout the entire section. Step 4: Perform topology optimization on the two-layer composite model, and remove auxiliary surfaces and non-manifold topologies generated during modeling through Boolean operations; based on anatomical landmarks, directionally remove the closed solids at the upper end of the esophagus and the distal end of the pylorus, and connect the entire digestive tract lumen to generate a single watertight, fully connected, uniformly thick cavity geometric model. Step 5: Based on the hollow solid, a finite element biomechanical simulation model is built. The displacement boundary conditions are used to simulate the three physiological narrowings of the esophagus without modifying the solid geometry to avoid mesh distortion. The Yeoh hyperelastic constitutive model is selected to characterize the nonlinear large deformation characteristics of soft tissue. A partitioned unidirectional coupling solution framework is used. The fluid domain in the stomach is set up with transient solution analytical unsteady flow, and the esophageal solid domain is obtained by steady-state solution to obtain the equilibrium stress field. The fluid excitation is transferred through unidirectional load mapping at the fluid-structure interface to complete the stable and convergent fluid-structure coupling numerical calculation.

2. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 1, characterized in that, The contour acquisition process in step 1 is as follows: the complete outer contour of the stomach is extracted from coronal slices and the complete outer contour of the esophagus is extracted from axial slices. The two sets of contour control points are merged and archived after unifying the spatial coordinate system to form a complete gastroesophageal contour dataset. The cubic B-spline curve used is generated by weighting each discrete control point with basis functions, which can smoothly fit the anatomical boundary with large curvature and suppress contour jagged distortion.

3. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 1, characterized in that, The surface smoothing and bidirectional shelling process in step 3 includes three optimization logics: First, G2-level surface continuity is achieved through dual constraints of curvature and torsion, preserving the key geometric features of the three physiological stenosis points while eliminating local concave and convex noise; Second, equidistant surfaces are generated by synchronous offset along the inner and outer normals of the model, which differs from traditional unidirectional shelling and avoids sudden changes in wall thickness and surface molding defects in high curvature areas such as the greater curvature of the stomach and the pylorus; Third, anatomical zones are separately divided for the outer and inner wall surfaces, and dedicated contour boundaries are added for the three physiological stenosis points to provide a precise geometric reference for subsequent mechanical boundary loading.

4. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 1, characterized in that, The process of topology cleaning and lumen connection in step 4 is as follows: First, remove the auxiliary patches that have no physiological significance generated during the modeling process to eliminate topological conflicts caused by multi-source geometric fusion; then, accurately remove the closed structures at both ends of the lumen to fully restore the physiological morphology of the continuous channel of the digestive tract, divide the geometric boundaries of independent fluid domains and solid domains, and meet the basic requirements of finite element mesh generation.

5. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 1, characterized in that, The logic for implementing the equivalent displacement boundary in step 5 is as follows: a displacement field matching the squeezing effect of surrounding tissues is applied to three physiologically narrow regions to dynamically replicate the periodic mechanical constraints of the swallowing process, completely avoiding the problems of local mesh distortion and element distortion caused by modifying the solid geometry.

6. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 1, characterized in that, The step 5 partitioned unidirectional coupling strategy: the fluid transient solution captures the time-varying pressure and shear flow field of gastric juice, and the solid steady-state solution obtains the esophageal equilibrium stress and strain state. Only the fluid calculation results are unidirectionally transferred to the solid mesh, without bidirectional iterative interaction, which greatly reduces the probability of calculation oscillation and iterative divergence caused by the superposition of the double nonlinearity of fluid and soft tissue.

7. The method for reverse modeling and convergent fluid-structure interaction simulation of the gastroesophageal tract with uniform wall thickness according to claim 6, characterized in that, The simulation material uses the Yeoh hyperelastic strain potential energy model to match the near-incompressible and large deformation mechanical characteristics of soft tissue. The fluid mass, momentum and energy conservation equations and the deformation gradient tensor are combined to complete the fluid-solid field coupling numerical solution.

Images