Design and performance study of biomimetic bone scaffolds based on the fusion of triply periodic minimal surfaces Authors: Yadi Sun, Yan Wang, Chunhui Ji, Yiyang Li, Jianxiong Ma Published: 2025 ABSTRACT AbstractAt present, the treatment of osteonecrosis of the femoral head (ONFH) with bone scaffolds has become a research hotspot. However, comparative studies on the performance of different fusion bone scaffolds are relatively scarce. This study adopted a method with a constant porosity of 70%. By adjusting the TPMS function equation, the three basic structures of D, G, and P were fused in pairs to form six fusion structures: D-G, G-D, D-P, P-D, G-P, and P-G. The simulations were performed using Ti6Al4V as the scaffold material, and through quasi-static compression experiments, we investigated the mechanical properties of these fused TPMS bone scaffolds. At the same time, we also used computational fluid dynamics (CFD) technology to study the effects of permeability, pressure, and flow rate on the fusion TPMS bone scaffolds. The results show that the fusion structure P-D has a low elastic modulus and a yield strength similar to that of natural bone tissue (153.6 MPa). The average wall shear stress of the P-D structure is 0.0037 Pa, the area suitable for cell growth accounts for as high as 98.9%, and it has the highest permeability (5.7245 × 10− 8m2).The design of the fusion porous scaffold meets the implantation requirements of human bone tissue.Keywords: ONFH, Biomimetic, Bone scaffolds, TPMS Introduction Osteonecrosis of the femoral head (ONFH) is an orthopedic disease caused by insufficient blood supply, which leads to bone necrosis and subsequent hip pain and dysfunction [1]. Studies have shown that femoral head collapse is a direct consequence of decreased mechanical properties and, if not treated early, most patients will collapse within three years [2]. Therefore, there is an urgent need to develop effective early ONFH treatment strategies to maintain hip function, prevent femoral head collapse, and reduce patients’ dependence on long-term treatment [3]. The current treatment methods are mainly divided into non-surgical treatment and surgical treatment. Non-surgical treatment mostly adopts conservative treatment methods, which can relieve pain and delay the course of the disease, but cannot prevent femoral head collapse [4]. Surgical treatments include total hip replacement and core decompression, but these operations are complex, traumatic, and have many complications [5]. Therefore, conservative treatment that can preserve the contour and mechanical properties of the femoral head, prevent joint collapse, and avoid arthroplasty is particularly important [6]. In recent years, studies have shown that implanting bone scaffolds into necrotic bone tissue is an effective treatment method. Bone scaffolds can accelerate and stabilize bone repair by providing rapid, robust, and durable mechanical support, thereby reducing the need for repeated interventions and significantly slowing the progression of early-stage ONFH [7]. The use of a bone scaffold with a porous structure significantly reduces the elastic modulus of the implant, avoids the stress shielding effect, and can provide good mechanical support for the femoral head [8, 9]. Furthermore, the interconnected pore network facilitates cell adhesion, proliferation, and tissue ingrowth, while enabling efficient nutrient and metabolite exchange with the surrounding tissues, thereby improving long-term biostability [10]. The structural diversity of bone scaffolds leads to heterogeneity in their mechanical and biological properties, which in turn show different responses in the interaction with the host bone after implantation. Therefore, how to construct bone scaffolds that meet the requirements is one of the research focuses. At present, in order to imitate the structure of natural bones, computer-aided design (CAD) technology is usually used, including regular structures, irregular structures and triply periodic minimal surfaces (TPMS), to design bone scaffolds that better meet performance requirements [11]. Although scaffolds with regular geometries are easy to fabricate, their sharp edges hinder integration with host bone tissue. In contrast, irregular structures offer greater geometric freedom but are difficult to precisely control in terms of pore morphology and uniformity [12]. Among them, the structure of TPMS is highly similar to the topological characteristics of human trabecular bone and can be described as a periodic infinite framework with zero average curvature at any point along three independent axes. Its superior biomimetic properties make it a popular choice for simulating natural bone structure [13, 14]. Compared with other grid-based lattice structures, TPMS is a periodic geometric structure with a lower level of stress concentration [15]. Structural diversity within TPMS scaffolds can be achieved by varying their implicit mathematical functions. Among them, the Primitive (P), Gyroid (G), and Diamond (D) types are the most extensively studied and widely applied [16, 17]. Most researchers tend to compare the performance of these homogeneous TPMS structures, but by fusing them into a dual-layer structure that mimics human bone, their performance can be optimized and different structures can be used in different areas [18–20]. Despite this, there is still a relative lack of systematic research on the structure of fused TPMS [21]. Natural bone tissue is a double-layer mixed structure, which contains two parts: a dense outer structure of cortical bone and a sparse inner structure of cancellous bone [11]. Bone exhibits multiscale structural diversity. At the macroscopic level, cortical bone provides high mechanical strength for load bearing, whereas cancellous bone features a porous architecture that supports metabolic exchange and shock absorption. At the microscopic level, cancellous bone consists of a trabecular network formed by interconnected struts, which ensures both mechanical adaptability and biological functionality. The structural characteristics of natural bone itself provide a template for constructing artificial bone scaffolds. The double-layer bone scaffolds designed based on biomimetic structures inspired by natural structures have better performance than homogeneous bone scaffolds [22]. Currently, uniform bone scaffolds cannot correctly mimic the human bone structure to assume different functions. After being implanted into the human body, under load conditions, they may cause excessive mechanical properties that cannot meet the performance requirements of the bone [23]. The double-layer scaffold design mimics the double-layer structure of natural bone, matching its mechanical and biological properties with the surrounding tissue at the implantation site. The cortical layer of the scaffold provides load-bearing strength, whereas the cancellous layer, with its porous network, facilitates nutrient exchange, supports cell proliferation and tissue ingrowth, and satisfies metabolic demands [24]. By adjusting the parameters of the double-layer structure, the internal lattice structure configuration can be adjusted to change the mechanical and biological properties of the bone scaffold [25]. Therefore, this study deeply explored the combined effects of different TPMS bone scaffolds on effective elastic modulus, yield strength, permeability, and wall shear stress through comprehensive numerical analysis. A hybrid TPMS design strategy was employed to replicate the dual-layer composite structure of cortical and cancellous bone. Six different double-layer hybrid TPMS bone scaffolds were designed by mixing three TPMS structures—P, G, and D—in pairs. Using numerical simulation technology, we compared the performance differences between these bone scaffolds to provide a scientific reference for the design of bone scaffolds. Materials and methods TPMS structural design method The structure of TPMS is a topological structure that can be controlled by mathematical functions. This paper uses implicit trigonometric functions in Matlab to approximate the function surfaces of P, G and D configurations. The above configuration equations correspond to the following [26, 27]: where x, y, z are spatial coordinates, ω = 2π/l, and l is the length of the unit cell. By setting a 2π periodic domain, a unit cell with constant porosity is simply formed, and the domain is repeated along the three axes of the Cartesian coordinate system to form the entire structure, as shown in Fig. 1. c is a constant, and by changing c, the normal direction of the isosurface can be controlled, and the volume of the subdomain can be changed [20]. Schematic diagram of the uniform TPMS model with (P), G, and (D) structures When this constant is equal to zero, the resulting minimum surface divides the space into two subdomains of equal volume. Modifying the level set constant, whether increasing or decreasing its value, will move the level set surface away from zero in the normal direction of the minimum surface. Therefore, the volumes of the two subdomains are adjusted accordingly, and the resulting uniform TPMS model is shown in Fig. 1. Double-layer axial fusion bone scaffold design In this study, we adopted a pairwise fusion strategy of three different TPMS lattice types to achieve structural and properties resembling those of cortical and cancellous bone. The bone scaffold model consists of two distinct structural regions—upper and lower—requiring the integration of different TPMS types. The double-layer bone scaffold is designed to mimic the natural architecture of bone, with an upper cortical layer and a lower cancellous layer. This configuration establishes a biomechanical gradient resembling that of human trabecular and cortical bone. This paper uses a weight function to fuse different structures, which helps to avoid stress concentration areas in the transition area, assign different TPMS structures to different subdomains in the model, and make different TPMS structures transition to each other and merge into one model [21, 28]. The fusion model of two TPMS structures can be defined as shown in Equation [29]. where represents the function of the final fusion structure, and and represent different units in the structure. is the spatial weight function, which can be described by the Sigmoid function, as shown in the formula [30]. where G(x, y,z) is the shape describing the transition between different regions, and k is inversely proportional to the width of the transition region. Previous studies have shown that the k value mainly affects local stress concentration, with a typical range of 1 to 4, and its variation will not have a significant impact [31]. The impact of changes in the k value on the sigmoid function is shown in Fig. 2a, and Fig. 2b shows that the width of the transition region is controlled by the k value. Therefore, in this study, in order to avoid stress concentration caused by transition mutations between structures, k is set to 1. In this study, three TPMS structures were mixed in pairs to build six bone scaffold models, using a 5 × 5 × 5 array unit cell, thinking that this could characterize the mechanical properties of an infinite array [32]. The upper two layers of cells represent the cortical bone region, the middle layer of cells serves as a transition zone, and the lower two layers of cells correspond to the cancellous bone. Its porosity is controlled at 70%. The length, width and height of the entire model are 30 mm, as shown in Fig. 3. a The impact of k value: Sigmoid function under different k values; b The area between the two red lines represents the width of the transition region in the fusion structure at different k values The flow chart of bone scaffold construction in this study shows a biomimetic strategy that combines the designs of three TPMS bone scaffolds based on G, D, and P to form a double-layer structure. This strategy aims to achieve a biomechanical gradient similar to that of human trabecular bone and cortical bone. a The left side shows the femoral head, with the yellow frame distinguishing the upper cortical bone and the lower cancellous bone, and the right side shows the constructed bone scaffold model; b Front view of the hybrid double-layer TPMS bone scaffold Mechanical performance simulation In this study, the commercial finite element analysis software ABAQUS was used to simulate the compression of bone scaffolds. A controlled load with a displacement of 30% (9 mm) was applied to the top surface to simulate physiological loading conditions in the human body, while the bottom surface was fixed. Contact conditions between scaffold surfaces were defined as frictionless to ensure stability during simulation, and the boundary conditions were shown in Fig. 4. The scaffold was meshed using 3D linear hexahedral elements (C3D8R) with an average element size of 0.3 mm, ensuring sufficient resolution for capturing stress and strain distributions. The material properties were configured based on Ti6Al4V: density of 4.41 g/cm³, yield strength of 944 MPa, elastic modulus of 110 GPa, and Poisson’s ratio of 0.3 [33]. Ti6Al4V was chosen as the scaffold material due to its excellent biocompatibility, high strength-to-weight ratio, and well-established use in orthopedic implants. During the simulation, the analysis steps were dynamic and explicit. The stress contour plot of the bone scaffold under 1% compressive strain was recorded. The force-displacement curve of the scaffold during compression was obtained through the Abaqus process output and converted into a stress-strain curve to obtain the effective elastic modulus and yield strength. a Computational domain for mechanical simulation with compression loading. Two rigid planes are applied at both ends of the domain, one of which is fixed and the other moves to apply displacement compression.b Schematic diagram of applying boundary conditions for fluid simulation Computational fluid dynamics simulation Permeability affects the transport of nutrients in the scaffold, which is essential for cell growth, while wall shear stress provides mechanical stimulation that can be sensed by cells, leading to better differentiation and proliferation [34, 35]. Fluid dynamics (CFD) analysis was performed using the commercial software COMSOL software to reveal the flow behavior of six double-layer hybrid bone scaffolds, and the permeability and wall shear stress of the bone scaffolds were evaluated. CFD modeling was approximated by single-phase and creeping flow models. In order to simplify the simulation calculation and analysis, the deformation of the metal scaffold during fluid flow was ignored. In this study, the analysis object is considered to be an incompressible fluid with constant density, which is defined by the Navier-Stokes equation, as shown in Eq. 6 [36]. where ρ is the fluid density (kg/m3), v is the fluid velocity (m/s), t is the time (s), µ is the dynamic viscosity coefficient of the fluid (kg/m·s), F is the force, ∇ is the operator, and P is the pressure (Pa). Figure 4b illustrates the computational domain and boundary conditions used for the CFD analysis. The flow direction is perpendicular to the inlet surface, the inlet velocity is constant at 0.001 m/s, and symmetric boundary conditions are used on the side walls. Zero gauge pressure is applied at the outlet, and the structural wall has a no-slip condition. In the fluid simulation analysis, the fluid Dulbecco’s modified Eagle’s medium (DMEM) was simulated as incompressible water, and the following fluid properties were obtained at 37 °C (average body temperature): its dynamic viscosity is u = 0.00145 Pa*s, and its density is 1000 kg/m³ [37, 38]. Based on the simulated flow behavior, the permeability and wall shear stress can be determined accordingly. The permeability K of the structure is determined by Darcy’s law, as shown in Eq. 7 [39]. The fluid-induced wall shear stress τw is determined from the product of shear rate and dynamic viscosity as shown in Eq. 12 [40]. where K represents the permeability (m2), v represents the fluid velocity (mm/s), µ represents the dynamic viscosity coefficient of the fluid (MPa ⋅ s), L represents the model length (mm), and ΔP represents the pressure difference (MPa). where τw is the wall shear stress (Pa), ˙γ is the shear rate (s − 1), and ˙εij is the surface velocity gradient tensor (s−1). Cell differentiation analysis In order to further study the differences in bone tissue differentiation among six double-layer fusion bone scaffolds, this paper conducted a computational analysis of the tissue differentiation of the bone scaffold wall under in vitro perfusion system culture conditions. In the in vitro culture environment, tissues differentiate along the inner wall of the bone scaffold, and this process is mainly affected by the load stimulation at the fluid-solid interface. Therefore, this study focuses on the tissue differentiation at the fluid-solid interface in order to reveal the mechanism of the influence of different bone scaffold structures on tissue differentiation. The mechanical stimulation of BMSC differentiation includes OSS and WSS. In order to analyze the different morphological scaffolds under the dual mechanical stimulation of fluid flow and strain, the total mechanical stimulation (SS) of different morphological bone scaffolds was quantitatively analyzed by FSI [41]. Prendergast [42] et al. proposed a mechanical regulation theory in which the fate of mesenchymal stem cells is regulated by the combined biophysical stimulation of tissue shear strain and fluid flow. The researchers used this theory to predict cell differentiation phenotypes. The mechanical stimulation indicators on the surface of the bone scaffold cover the surface OSS of the solid domain and the WSS of the fluid-solid interface. The OSS is calculated based on the solid domain model, while the WSS is calculated using the fluid domain model. After completing the above calculations, the mechanical regulation stimulation value of each bone scaffold is further calculated to predict the cell differentiation inside the bone scaffold. The specific calculation formula is: where SS is the mechanical stimulation factor, OSS is the octahedral shear strain, WSS is the wall shear stress, and the constants a and b are 0.0375% and 10 mPa. According to the modified mechanoregulation theory, tissue phenotypes are predicted by calculating stimuli and their classification: If SS ≤ 0.01, the stimulus is too low and no tissue differentiation is predicted; If 0.01 < SS ≤ 1, bone tissue differentiation is predicted. If 1 < SS ≤ 3, cartilage tissue differentiation is predicted. If 3 < SS ≤ 6, fibrous tissue differentiation is predicted. If SS > 6, the stimulus is too high and no tissue differentiation is predicted. The OSS is calculated according to the fluid-solid coupling model as follows: where 、、 are the elastic principal strains, defined as: where, is the principal strain in the i-th direction, and is the principal stretch ratio in the i-th direction. The WSS on the surface of the bone scaffold was calculated according to the laminar CFD model as follows: where is the dynamic viscosity and is the normal velocity gradient of the fluid at the wall. In order to calculate SS, the two output parameters OSS and WSS must be calculated at the same element nodes [41]. During meshing, adjacent elements on the contact surface share nodes, ensuring continuity of physical quantities (e.g., displacement and stress) and enhancing the accuracy and efficiency of the calculations. During the mesh generation process, special attention is paid to the surface of the bracket, because this area needs to be properly refined to calculate the mechanical adjustment force. As can be seen from Fig. 5, the conformal mesh technology is applied between the solid and fluid environments on the surface of the bracket. The result is that the structural and fluid properties can be calculated at the same node. A mesh sensitivity analysis was previously performed to understand the impact of mesh sensitivity on the simulation results. Therefore, the meshing method in this section is consistent with that described above, and the entire bone bracket structure and fluid domain are spatially discretized. a Conformal mesh between fluid and solid regions applied to the bone scaffold surface; b Schematic diagram of the fluid-structure coupled cell differentiation process The solid domain was used to simulate the uniaxial compression loading experiment under bioreactor conditions, while the fluid domain was used to simulate the fluid flow under perfusion culture conditions. The specific meshing parameters remained the same as before. The substrate of the solid domain (bone scaffold) FE model was Ti6Al4V (elastic modulus E = 110 GPa, Poisson’s ratio = 0.3, and density 4.41 g/cm³). The lower surface was constrained in all directions. A compressive load of 0.25 N/mm2 was applied to the upper surface, replacing the original 2% strain load, to simulate the low to medium physiological loads acting on cancellous bone [43]. The OSS of all models was calculated, and the extraction of WSS followed the same method described in the CFD simulation. When calculating the mechanical stimulation factor, the alignment between the mesh nodes of the bone scaffold surface in the compression FE simulation and the CFD model was maintained. Results Mechanical properties of fusion bone scaffolds In the field of modern orthopedic repair, the design and application of bone scaffolds are of vital importance, especially in providing mechanical support and promoting bone tissue regeneration. This study shows in detail the stress distribution of six double-layer hybrid TPMS bone scaffolds with different designs under constant deformation pressure through Fig. 6. In the experimental setup, we used a precisely controlled constant deformation pressure to simulate the mechanical environment that bone scaffolds may encounter in actual applications. The six bone scaffolds, including P-G, G-P, P-D, D-P, G-D and D-G, all showed their own unique stress response patterns. Stress contour plot of bone scaffold As shown in Fig. 7, during the compression process, for the P-G and G-P bone scaffolds, the stress concentration phenomenon was significantly larger than the other four structures, and the stress concentration mainly appeared on the P structure, and these stress concentration areas were perpendicular to the direction of the applied external force. For the P-D and D-P bone scaffolds, although they also showed similar stress distribution trends, it is worth noting that their maximum stress values ​​were significantly lower than those of the P-G and G-P structures. This may mean that these bone scaffolds have greater safety and stability when subjected to the same pressure. Further observing the G-D and D-G bone scaffolds, we found that stress concentration areas mainly appeared on the G structure, and these areas were parallel to the direction of external force. This stress distribution characteristic is highly consistent with the force behavior of bones, indicating that the design of G-D and D-G bone scaffolds is more in line with the biomechanical properties of human bones. That is, the compression behavior of G-D and D-G bone scaffolds is more consistent with the natural stress pattern of human bones. This bionic design not only provides better mechanical support, but also helps promote the natural growth and repair of bone tissue. Therefore, these findings are of great significance in guiding clinicians to select appropriate bone scaffolds and provide patients with more effective treatment options. Stress contour plot on the front of the bone scaffold Elastic modulus and yield strength are important indicators for evaluating the mechanical properties of porous structures, which can measure the ability of the structure to resist elastic deformation and withstand external loads. As a load-bearing part that replaces the natural bones of the human body, the porous structure must have a certain strength to ensure that it is not damaged under the action of external loads. Figure 8 shows the comparison of the elastic modulus and yield strength of 6 porous structures. Among the 6 porous structures: the elastic modulus (19.906GPa) and yield strength (178.8 MPa) of the DG structure are the largest, and the elastic modulus (12.474GPa) and yield strength (142.3 MPa) of PG are the smallest. The elastic modulus and yield strength (142.3–178.8.3.8 MPa) of the six structures are within the requirements of implantable prostheses for the human body. The elastic modulus of natural cortical bone is reported to be between 7 GPa and 30 GPa, while its yield strength is typically in the range of 100 MPa to 200 MPa [44]. These mechanical properties ensure adequate load-bearing capacity and compatibility with surrounding bone tissue, minimizing the risks of implant failure or stress shielding effects. sicouu :a Elastic modulus for different scaffold designs; b Yield strength for different scaffold designs Permeability of fusion bone scaffold Nutrients such as blood, oxygen, and cells in the human body will circulate in the internal channels of the scaffold, and new bone tissue will also grow from the outside of the scaffold to the inside of the scaffold along the channels of the scaffold. Therefore, for better growth of bone tissue, it is necessary to ensure adequate transmission of nutrients and cells within the porous channels. Therefore, it is very helpful to explore the flow pattern of nutrients in the scaffold pores and the permeability of the designed structure for bone tissue regeneration. Figure 9 shows the fluid velocity contour plots of six types of axially fused porous scaffolds. In the six structures, the flow rate of the flow channel near the fusion point of the scaffold is significantly higher than that of the surrounding area, which also ensures that nutrients can be transported into deeper pores and avoids the difficulty of new bone tissue growing into the interior of the scaffold. problem, it also confirms that the axial fusion design is conducive to the flow of nutrients and cell migration; while the flow velocity in the peripheral area is smaller, this is mainly due to the greater frictional contact between the fluid and the hole wall in the peripheral area, and the smaller flow velocity and larger Friction can give cells more time to adhere to the surface of the porous structure, thereby increasing the cell adhesion rate. In addition, the fluid velocity of the D-G and G-D structures is higher, and the flow velocity of the porous structure is evenly distributed. This is mainly attributed to the regular distribution of pore channels and small pore diameters in the G and D structures; the G-P and P-G structures have a large number of large pores due to the P structure and the pores are distributed axially, causing the fluid flow rate to be small and the flow rate to accumulate at the P structure. Bone scaffold flow velocity contour plot The permeability of all stents was calculated according to Eq. 7, and the results are shown in Fig. 10. The two fusion structures of D and P structures can achieve a larger coverage area of 5.26 < K(×10− 8m2) < 5.72 than the other four regular porous scaffolds, in which D-P permeability: 2.4602 × 10− 8m2 and P-D permeability: 5.7245 × 10− 8m2. The decrease in permeability is mainly due to the large resistance faced by the fluid with small D pores in the upper layer. As the P pores in the lower layer increase, the resistance decreases, and the permeability of the scaffold begins to gradually increase. What is interesting is that the permeability of the two structures G-P (5.13 × 10− 8m2) and P-G (5.34 × 10− 8m2) is close. This is mainly due to the fact that the G and P structures themselves have better fluid circulation. Even if Changing the fusion positions of G and P will not significantly change the permeability. However, for the bionic design of cortical-cancellous bone, the same permeability difference as the D and P structures is required. Overall, the permeability of the six structures is within the range of human bone permeability (0.5 < K(×10− 8m2) < 6.0), which is similar to the results of Castro et al.‘s research and can meet the usage requirements [45]. a Permeability under different scaffold designs; b Variation of average wall shear stress under different scaffold designs Wall shear stress of fusion bone scaffold Another important feature of the fluid response of the scaffold is the fluid wall shear stress (WSS). The liquid-induced FSS will provide mechanical pulses, which are sensed by cells and lead to better cell proliferation and differentiation. When the fluid wall shear stress is between 0.05 and 57 mPa, it is conducive to cell growth; when the shear stress is less than 0.05 mPa or greater than 57 mPa, cell growth is inhibited [37]. The WSS distribution contour plot on the inner wall of the scaffold is shown in Fig. 11. For all models, the calculated average WSS varies between 3.08 and 9.65 mPa, which is similar to the research results of Derya Karaman et al. [46]. Combined with Fig. 10, it is found that the average wall shear stress structure of the G-D structure is due to its large specific surface area, which is mainly due to the high degree of tortuosity of the G structure wall, the flow velocity rotation is accelerated, and a large wall shear stress is formed. The average wall shear stress of the two fusion structures of the D and P structures is not much different, and is within the design requirements of the wall shear stress. After calculation, the average wall shear stress of the six axial fusion structures was in the range of 0.1–10 mPa, and the results were similar to those of Kumar and Tang et al. [47] Fig. 12 shows the proportion of suitable cell growth area. Among them, the proportion of P-D suitable for cell growth area is 98.2%, which is much larger than that of other scaffolds. From this point of view, the fusion scaffold P-D has a better wall shear stress distribution and is more suitable for the attachment of bone cells after implantation. Contour plot of shear stress on the bone scaffold wall Calculation of cell differentiation area based on wall shear stress Cell differentiation Due to the changes in the bone scaffold structure and the restriction of fluid flow in different regions, tissue differentiation in the bone scaffold is uneven. Computational models have been applied to predict the mechanical stimulation experienced in the bone scaffold under in vitro flow perfusion or mechanical compression. As can be seen from the SS distribution cloud map in Fig. 13, the stimulation intensity of bone scaffolds with different structures shows significant differences under the same strain and inlet flow rate conditions. In all bone scaffolds, there are regions with high and low SS values, which are closely related to the distribution of wall shear stress and strain. High SS regions usually correspond to high WSS and high strain regions, which are often concentrated in areas with more active fluid flow. These regions have high WSS and strain values, which are conducive to cell proliferation and differentiation. Low SS regions correspond to low WSS and low strain regions. Low SS regions often appear in areas with limited fluid flow or denser structures. These regions have low WSS and strain values, and cell differentiation may be inhibited. The SS distribution of G-P and P-G structures is relatively more uniform, and the high SS region is more widely distributed. This is mainly due to the relatively tortuous internal shape of the G structure, which has a significant dispersion effect on fluid flow, making the WSS distribution more uniform. This uniform WSS distribution indicates that the G-P and P-G structures have obvious advantages in mechanical stimulus transmission, which makes the differentiation in osteocytes and chondrocytes more uniform. Compared with other structures, the SS distribution of P-D and D-P structures is more concentrated, and the high SS area is mainly concentrated in the transition zone of the two structures. This phenomenon may be due to the existence of a direct current channel inside the P structure, which causes the fluid to directly impact the transition zone, thus forming a local high SS value. Although this concentrated distribution may affect the uniform differentiation of cells, high SS values ​​in specific areas contribute to the differentiation of specific types of cells, such as osteocytes. Among the six bone scaffolds, the P-D and D-P structures have the highest degree of osteocyte differentiation. The SS distribution of D-G and G-D structures is between the above two situations, showing both a certain uniformity and a certain concentration. This may be because the internal shapes of the D and G structures are relatively tortuous, which has a dispersive effect on the fluid flow, making the WSS distribution relatively uniform. In general, there are significant differences in the SS distribution of bone scaffolds with different structures. These differences are mainly attributed to the structural design of the bone scaffold and the restriction of fluid flow, which provides an important theoretical basis for the design of bone scaffolds. SS distribution on the surface of bone scaffold As can be seen from Table 1; Fig. 14, under the model and conditions defined in this study, all bone scaffolds achieved more than 99% osteocyte and chondrocyte differentiation. The proportion of BMSCs that differentiated into fibroblasts or failed to differentiate was extremely small. Compared with OSS, mechanical stimulation factors are more affected by the WSS value. Studies have shown that the double-layer fusion structure designed based on TPMS can effectively promote the differentiation of BMSCs into osteoblasts under mechanical stimulation. Specifically, the bone scaffold with the P-D structure showed the best osteogenic differentiation effect, with the osteoblast differentiation rate reaching 99.36%. This phenomenon shows that the P-D structure not only has good stress transfer characteristics, but also provides a good microenvironment for cells, thereby promoting the formation of bone tissue. Compared with the P-D structure, although the D-P structure also showed a higher effect in osteogenic differentiation, with a ratio of 99.21%, it further differentiated into more chondrocytes (0.69%) and fibroblasts (0.09%). The P-D structure had only 0.49% chondrocytes and 0 fibroblasts, which may be due to the relatively uniform stress distribution of the D-P structure and the lack of local high stress areas to promote cell proliferation. In contrast, the cell differentiation effect of the D-P structure is weaker, mainly due to the stress concentration of the bone scaffold, which leads to the restriction of cell attachment and growth. The D-G and G-D structures show similar cell differentiation trends, moderate osteogenic differentiation effects, and their differentiation ratios are slightly lower, 99.40% and 98.83% respectively. The osteoblasts of the D-G structure are slightly higher, but the other two are more differentiated. However, the promotion effect of differentiation on cell differentiation is still obvious, and the cell growth space is effectively guaranteed. The osteoblast differentiation ratios of the G-P and P-G structures are the lowest, 96.90% and 98.24% respectively, but the two structures have the highest differentiation of chondrocytes among the six structures, 1.08% and 1.12% respectively. This phenomenon suggests that in some structures, the interaction between fluid shear stress and dynamic strain of the bone scaffold may play a positive role in promoting the differentiation of chondrocytes. The internal structure of the G structure is relatively tortuous, especially in the area where the fluid flow is more active in the scaffold. Table 1Quantitative distribution of differentiated tissue types on the surfaces of six fusion bone scaffoldsUndifferentiatedOsteocytesCartilageFibroblastsSuper DifferentiationSS≤0.010.01