Lightweight Knee Orthosis for Athletic Rehabilitation: Achieving 40% Weight Reduction with Topology Optimization in Generative AI Design

1. Introduction

This proposed work focuses on developing a lightweight and customizable knee orthosis tailored for general rehabilitation and athletic recovery using Functional Generative Design (FGD) on the 3DExperience platform. Traditional off-the-shelf knee braces often lack personalization, leading to bulky and uncomfortable products that can hinder usability and cost-efficiency. By leveraging recent advancements in generative design and additive manufacturing, this study aims to create a structurally optimized, user-friendly brace through iterative simulations, prototyping, and refinement. Although anatomical data was collected using 3D scanning, the design was intentionally generalized to test broader usability beyond a single user’s profile, enhancing its application in diverse rehabilitation settings. Incorporating clinical insights, the design process aligns with evidence-based recovery benchmarks, particularly targeting a knee flexion range within the first month post-surgery. This parameter serves as a design requirement for the hinge mechanism to ensure sufficient early-stage support. Beyond the technical development, the proposed work showcases how engineering, healthcare, and digital manufacturing intersect to offer scalable, personalized solutions in orthopedic care.

2. Literature Review

Generative design and topology optimization are increasingly used to develop orthotic devices that minimize weight while maintaining structural integrity. In biomedical applications, these methods enable the creation of complex, material-efficient geometries without compromising strength or functionality. Custom wrist splints validated by finite element analysis have achieved high motion restriction with low discomfort, supporting dynamic rehabilitation [1]. In prosthetic foot design, compliant mechanisms replicate natural gait through controlled deformation, reducing material usage compared to rigid structures [2]. Lattice-optimized brace shells balance support and airflow, enhancing long-duration comfort [3], while anatomically tailored prosthetic sockets improve stress distribution using CATIA-based topology workflows [4]. Algorithmic refinement of lattice structures enables designers to trade stiffness for reduced weight, improving personalization and fatigue resistance [5]. Simulation-driven optimization integrated into CAD platforms has further refined jointed ankle supports by directly incorporating gait mechanics [6]. Collectively, these approaches can reduce material usage by up to 50% while preserving structural performance, making them well-suited for lightweight athletic orthoses. Lower-limb orthoses must accommodate biomechanical loads and dynamic gait patterns. Studies have shown that intentional hinge misalignments in braces can alter joint kinematics but may not always reduce ligament strain, emphasizing the importance of anatomical alignment [7]. Valgus and external rotation configurations, especially when combined with lateral wedge foot orthoses, significantly reduce medial knee loading and improve comfort [8]. However, excessive rigidity in ankle-foot orthoses can suppress muscle activation, potentially hindering gait retraining and neuromuscular recovery [9]. Community programs using 3D-printed assistive orthoses have reported improvements in gait speed and user confidence, validating the role of adaptable structural design [10]. Improper strapping or misalignment during donning remains a critical failure point, often negating intended biomechanical benefits [6]. Postoperative studies on knee flexion after total knee arthroplasty (TKA) identified one-month flexion as a strong predictor of long-term recovery, with 105° at one month serving as the optimal benchmark for achieving 120° at 12 months [11]. Based on this, the target range of motion for this study was defined around 105° flexion, informing the hinge range of motion specifications. Aligning the design objectives with clinical recovery benchmarks ensures that the prototype provides effective structural support during early rehabilitation.

2.1. Methodology

The design process began with an initial attempt to import a 3D-scanned human leg model into the Functional Generative Design environment on the 3DExperience platform. However, due to limitations in handling complex mesh data, manual measurements were instead used to approximate the lower limb geometry. A semi-circular support structure was designed for the thigh and shin, combined with adjustable velcro straps to accommodate varying leg sizes. Critical design regions such as strap holes, bolt holes, and hinge connectors were preserved using partitioning tools, while the remaining areas were left open for material optimization to improve weight-to-strength performance. Once the design regions were defined, preserved zones and bolt connections were specified to ensure structural continuity during optimization. Structural loads simulating knee flexion forces were applied to the outer surface, while fixed constraints were assigned to the velcro-strap areas. Directional restraints controlled unwanted movement, and a mesh was generated to validate stress and deformation. Optimization parameters included a mesh size of 3.3mm with a minimum thickness control of 10mm, and symmetry enforcement to ensure durability and manufacturability. After defining a mass constraint, the topology optimization was executed, iteratively removing material until an optimal design balancing stiffness, weight, and form was achieved. The mechanical properties of PLA were applied in the simulations to determine its applicability for 3D-printing [12]. Base designs of a thigh and shin brace with connecting stems were first created in Figure 1, with original masses of 0.131kg and 0.199kg respectively.

The finalized digital designs were fabricated by 3D-printing with PLA. The use of additive manufacturing allowed for rapid iteration, particularly given the complex geometries resulting from topology optimization. While the generative forms were lightweight and efficient, printability challenges arose with thin overhangs and intricate interior cavities, requiring strategic support structures and orientation adjustments. Physical prototypes were tested on a user’s leg to assess fit, comfort, and alignment, validating the practical application of the design and informing future refinements.

2.2. Design Results

The shin and thigh components were designed with functional constraints to optimize weight, fit, and flexibility. Shin design focused on accommodating muscle expansion and ease of adjustment, using velcro strap holes and an open posterior. The thigh brace followed a semi-cylindrical shape to offer stable support while allowing leg movement. Both designs were connected by rotational hinges, engineered to support natural knee flexion using bolt-secured connectors. After the basic models were created, topology optimization was performed with defined cutting values (75 for shin, 70 for thigh) to remove unnecessary material while maintaining structural integrity. Simulation results showed significant stress reduction in optimized models, particularly in high-stress areas like strap holes. Experiments involved a remote force of 5000N being applied normally to the brace’s surface, shown in Figure 1 as green vectors, to simulate and assess the physical stresses exerted on a brace. The Von Mises stress results experienced by the braces were also shown in Figure 1, with the maximum displacements for both braces used in a comparative analysis in Table 1.

From Equation 1, The Von Mises stress σ_v provides an equivalent stress value that indicates whether a ductile material (PLA in this context) reaches the yield strength σ_y of the material under complex loading conditions.

$${\sigma }_v=\sqrt{{\displaystyle \frac{1}{2}}\left[\right({\sigma }_x-{\sigma }_y{)}^2+({\sigma }_y-{\sigma }_z{)}^2+({\sigma }_z-{\sigma }_x{)}^2]+3({{\tau }^2}_{xy}+{{\tau }^2}_{yz}+{{\tau }^2}_{zx})}$$

(1)

The simulations performed on the base model produced Von Mises stress values that remained within the safe limits of PLA’s mechanical capacity, with peak stresses observed at the velcro strap holes. The low stress values observed at the point of impact during the structural simulation indicated that the brace was able to evenly distribute the force around its entire frame, and did not exceed the typical yield strength of PLA [13]. This confirmed that under expected load conditions, the brace would not undergo plastic deformations or structural failure. The maximum stress can be shown in the four strap holes as the colour around the area is red and yellow. While the rest of the thigh design area experiences less stress. It is due to less reaction forces applying on the surface area when the user is wearing it to move around. Figure 2 illustrates the conceptual shapes of the shin and thigh designs, optimized through selective material removal. A cutting value of 70 was applied to the shin to define a clearer yet structurally valid shape, balancing weight reduction with integrity. For the slightly shorter thigh brace, a cutting value of 65 was used to preserve critical structural paths while minimizing excess material. Both thresholds produced lightweight, stable forms suitable for further analysis and prototyping without compromising functional support.

Figure 3 shows the same loads applied on the conceptual cases with a 50% mass constraint after optimisation. This mass constraint led to the optimised thigh brace having a new mass of 0.08kg, resulting in an approximate 0.051kg weight reduction; and the optimised shin brace having a new mass of 0.137kg, resulting in an approximate 0.062kg weight reduction, which helped to provide definite benefits in terms of wearer comfort and mobility. Because it lessens fatigue and improves the brace’s wearability, this reduction is especially important for sportsmen or recuperating patients performing repetitive exercises. Additionally, the 50% mass constraint models’ von Mises stress increased by approximately twofold according to finite element calculations, indicating a less effective stress distribution throughout the modified lattice structure. This was a surprising but positive result, showing that the generative method redistributed material under more stringent mass limitations, lowering critical stress concentrations without sacrificing performance as shown in Table 1 below

In wearable orthoses, higher displacement is undesirable as it can result in localized deformation during impact, potentially causing discomfort or injury if the brace presses into the user’s skin. Therefore, maintaining lower displacement values is essential to ensure both comfort and safety. Trade-off studies were tabulated in Table 1 for the thigh brace and shin brace, and mass constraints of 40%, 45%, and 50% were evaluated to understand the balance between weight savings and structural performance. Results from the original models were also tabulated in the same table. Trade off study scores out of 100 were used to determine the best mass constraints for both braces, with the same input load being used from Figure 1’s simulations. Mass constraints below 40% were avoided to prevent excessive material removal that could compromise the overall stiffness and structural integrity of the brace. While lower mass improves mobility and reduces fatigue, a minimum threshold is needed to preserve the mechanical robustness required for repeated dynamic use.

Following these results, a graph of the reduction in weight for both the base models and optimised models is shown in Figure 4.

Overall, although the 50% mass constraint design for the shin brace was slightly stiffer, it showed that higher stress levels could cause discomfort or shorten its lifespan when subjected to repeated loading. The 45% model for the thigh brace, on the other hand, is a better choice for dynamic use cases, such athletic therapy or later-stage recovery involving frequent mobility, as its minor weight decrease is balanced by its enhanced stress management and structural durability. The final design was 3D-printed with PLA and was physically tested on a user’s leg to validate fit and function. To connect the knee brace and shin brace together, a hinge was 3D-printed with a maximum angle of rotation of 105°. The brace shown in Figure 5 successfully aligned with anatomical landmarks, and the hinge system operated smoothly during flexion. The user reported comfort and adequate support without restriction, confirming the effectiveness of the topologically optimized structure. These results support knee orthosis as a viable, ergonomic solution for athletic rehabilitation and daily wear.

3. Discussion

The final knee brace design was selected following extensive simulation and trade-off analysis, optimizing for mass reduction, structural integrity, and stress tolerance. The thigh brace implemented a 45% mass constraint, while the shin brace adopted a 50% constraint, together resulting in a combined material weight of 0.208kg, a reduction in weight of 36.97% compared to the combined weight of the original model (0.330kg) in Figure 5. This reduced the total prototype weight by approximately 63 grams and significantly lowered material usage, yielding tangible cost savings in 3D printing. While designs with more aggressive material removal offered additional weight reductions, they introduced higher stress concentrations and excessive deformation, making them unsuitable for rehabilitation use. Conversely, less optimized models-maintained strength but failed to deliver the desired efficiency gains. Practical considerations such as hinge alignment and printability also influenced the final configuration. Although topology optimization generated high-performance geometries, certain features, such as thin overhangs and unsupported cavities, posed challenges during printing. These issues required minor geometric adjustments and careful print orientation to minimize support structure usage and reduce damage risk during post-processing. This iterative refinement process underscored the importance of integrating design-for-manufacturing principles early in development, especially when engineering orthotic components where material efficiency, cost, and structural performance must be carefully balanced.

4. Conclusion

This proposed work successfully demonstrated the use of Functional Generative Design in combination with 3D-printing to create a lightweight, structurally efficient knee brace for sports rehabilitation. The final prototype balanced material efficiency, mechanical performance, and user functionality, though challenges with geometry complexity and printability were noted. While the design met many objectives, future improvements, such as integrating patient-specific scans, exploring alternative materials, and conducting mechanical testing, could enhance customization and validation. The study highlights the potential of simulation-driven design and additive manufacturing for developing personalized, eco-friendly orthotic solutions. Using PLA supports sustainability, and expanding the brace to support the entire leg could further improve injury prevention and rehabilitation. Overall, this work showcases how combining generative design with material science can drive innovation in accessible and effective orthopedic devices.