🔬 condensed matter

Medial Model of Elastic Curve Contacts, from Clasps to Knits

This paper proposes using medial axis transformations and their radical generalizations to model the complex, saddle-shaped contact surfaces between interlocking fibers, demonstrating that these geometric methods effectively approximate contact mechanics in structures ranging from clasps to knits.

Original author: Kathryn Chain, Haoye Sun, Michael S. Dimitriyev

Published 2026-11-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original author: Kathryn Chain, Haoye Sun, Michael S. Dimitriyev

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the author. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world built not of solid blocks, but of tangled threads. From the sweater on your back to the complex nets that hold together the microscopic scaffolds of advanced materials, these structures rely on fibers weaving around one another. For centuries, scientists have understood how solid, round objects press against each other. When two smooth balls touch, they flatten slightly at a single point, and the force required to push them together follows a predictable rule. This understanding works well for simple shapes. But the fibers in a knit or a knot are different. They are long, flexible, and often curve back to wrap around each other, creating contact surfaces that are not flat or point-like, but saddle-shaped and deeply interlocked. These complex shapes have long been a blind spot in physics, making it difficult to predict how tangled materials stretch, compress, or hold their form.

A team of researchers at Texas A&M University has taken a fresh look at this problem, focusing on the hidden geometry where fibers meet. Instead of trying to model every tiny deformation of the material itself, they asked a simpler question: what does the surface look like where two fibers press against each other? They suspected that these contact zones might resemble a specific type of mathematical surface known as a minimal surface, a shape that naturally minimizes its area, much like a soap film stretched between two wire frames. To test this, they built a computer model of elastic fibers that could bend and stretch, simulating how they would settle into a stable position when tangled. They examined two specific scenarios: a simple cross where two identical fibers meet at right angles, and a basic knitted pattern where loops of yarn interlock.

The researchers found that their initial guess was only partially right. In the case of two identical fibers crossing each other, the contact surface did indeed look very much like a minimal surface in the very center, where the fibers pressed most firmly. However, as they looked slightly further out from that center point, the surface began to curve in ways that a minimal surface would not. The shape developed a distinct "saddle" form, curving up in one direction and down in another, which meant it was not perfectly flat in the way a soap film is. This deviation was significant; it suggested that while the idea of a minimal surface is a good starting point, it is not the whole story for how fibers actually touch.

The picture became even more complex when the researchers introduced a second variable: fibers of different thicknesses. In the real world, yarns are rarely perfectly uniform, and knitted fabrics often mix different types of thread. When they simulated a thin fiber wrapping around a thicker one, the contact surface shifted away from the center line between the two fibers. The standard geometric tools used to map these spaces failed to predict the true contact point. To solve this, the team developed a new method based on a concept called the "radical axis," which they adapted to account for the different sizes of the fibers. By adjusting the mathematical weights in their model to reflect the actual physical contact, they created a new approximation that matched the simulated contact points much more closely than the old methods.

This new approach revealed a surprising trend. When fibers of different sizes touch, the surface between them does not try to minimize its area to zero curvature. Instead, it settles into a shape with a specific, finite amount of curvature. The researchers observed that the contact surface seemed to prefer a state where the average curvature was reduced, even if it wasn't zero. This finding challenges the idea that these tangled structures always seek the simplest possible geometry. Instead, the interplay between the flexibility of the fiber and the pressure of the contact creates a unique, stable shape that is distinct from both the simple minimal surface and the flat contact of rigid objects.

The study also looked at how these shapes change when the material is stretched. When the researchers pulled on their simulated knitted fabric, the contact points shifted, and the curvature of the touching surfaces changed in a predictable way. The fibers concentrated their bending in the areas where they clasped together, while the rest of the structure remained relatively straight. This behavior helps explain why knitted fabrics are so stretchy and resilient; the fibers can slide and reconfigure their contact points without breaking, absorbing energy through these subtle geometric shifts.

Ultimately, this work provides a new map for understanding the hidden geometry of tangled matter. By moving beyond the old rules for how round objects touch, the researchers have shown that the contact surfaces between flexible fibers are dynamic and shape-specific. They have demonstrated that while simple geometric models can offer a glimpse into these interactions, a more sophisticated approach is needed to capture the reality of how fibers conform to one another. This new "radical axis" method offers a practical tool for predicting how these complex structures behave, paving the way for better designs in everything from high-performance textiles to advanced metamaterials that rely on tangled networks to function. The results suggest that the secret to the strength and flexibility of these materials lies not in the fibers themselves, but in the precise, curved geometry of the space where they meet.

Technical Summary: Medial Model of Elastic Curve Contacts, from Clasps to Knits

Problem Statement
Modeling materials composed of interlocking and tangled fibers, such as knitted fabrics, presents significant challenges due to the complex interplay between fiber geometry and contact mechanics. Traditional contact mechanics, rooted in Hertzian theory, describes interactions between convex bodies (e.g., spheres or cylinders) where contact occurs at a point or line. However, fiber structures like clasps and knits often involve non-convex, saddle-shaped contact surfaces where bodies conform to one another. This geometry falls outside the scope of classical Hertzian descriptions. Furthermore, while recent studies suggest these contact surfaces may resemble minimal surfaces (surfaces with zero mean curvature), the mechanical description of such "clasps" remains underdeveloped, particularly regarding stress transfer and the behavior of non-identical fibers with differing radii or stiffnesses.

Methodology
The authors propose a geometric approach to approximate contact surfaces between elastic curves without relying on full 3D continuum elasticity simulations. The methodology consists of three main components:

Elastica Model with Power-Law Contact: The authors utilize a modified 3D elastica model where fibers are represented as centerline curves with cross-sectional radii. Contact interactions are modeled via a power-law potential ( $u \sim \delta^{p+1}$ ), where $\delta$ is the deformation depth. For $p=3/2$ , this recovers Hertzian contact. The model treats the contact surface as a degree of freedom, minimizing total elastic energy (bending, stretching, and contact energy) to determine the deformed state and the precise location of contact points.
Medial Axis Transformation: To approximate the contact surface geometry, the authors employ the medial axis transform (MAT). This generalizes the Voronoi tessellation from points to curves. By discretizing fiber centerlines into vertices and generating a Voronoi tessellation of space, the union of cells associated with a specific curve defines a "medial region." The boundary of this region serves as an approximation of the contact surface between identical curves.
Radical Axis Transform for Asymmetry: To address cases where fibers have different radii (polydispersity), the authors generalize the MAT using "radical Voronoi tessellations." This method replaces Euclidean distance with a "power" metric ( $\Pi = d^2 - R^2$ ). Recognizing that the standard radical axis assumes non-deforming interpenetration, the authors introduce a "corrected radical axis." This involves calculating effective radii ( $R_{eff}$ ) for the generating curves such that the resulting radical axis passes through the actual contact points determined by the energy minimization model.

Key Contributions and Results
The study applies these methods to two specific geometries: an orthogonal clasp (two identical or non-identical curves crossing at 90 degrees) and a simple knit structure (stockinette stitch).

Symmetric Cases (Identical Fibers): In the case of identical curves (orthogonal clasp and simple knit), the medial axis transform provides a reasonable approximation of the contact surface. Curvature analysis reveals that while the center of the contact region approximates a minimal surface (zero mean curvature), significant deviations occur near the periphery. The contact surfaces exhibit saddle shapes (negative Gaussian curvature) but possess finite mean curvature gradients, challenging the hypothesis that these surfaces are strictly minimal.
Asymmetric Cases (Different Radii): When fibers have different radii, the standard medial axis fails to accurately predict the contact surface. The "corrected radical axis" transform, which adjusts effective radii to align with contact points, provides a significantly better approximation than both the standard medial axis and the uncorrected radical axis.
Curvature Trends: In asymmetric configurations, the contact surfaces exhibit non-zero mean curvature. The corrected radical axis consistently shows a slightly lower average mean curvature compared to the uncorrected version, suggesting an energetic preference for surfaces with reduced mean curvature, even in the absence of perfect minimal surface geometry.
Knit Structures: The method successfully extends to periodic knit structures. The medial axis of a relaxed knit forms a periodic handlebody with helicoidal patches where contact occurs. Stretching the knit increases the magnitude of Gaussian curvature in the clasp regions while maintaining the overall symmetry of the mean curvature distribution.

Significance and Claims
The paper claims to provide a direct link between fiber contact mechanics and geometric modeling methods. By leveraging the medial axis and radical axis transforms, the authors demonstrate a computationally efficient way to approximate contact surfaces between elastic curves. The results suggest that while contact surfaces in highly symmetric cases approach minimal surfaces, they generally adopt finite mean curvature, particularly when fiber asymmetry is introduced.

The authors position this work as a step toward more sophisticated geometric methods for modeling tangled metamaterials. They explicitly state that their results "may point the way" toward better models but acknowledge that the approach requires validation against experiments and finite element analysis. The paper modestly concludes that while the Hookean form is commonly used, the specific constitutive models for curve interactions remain an open question, and their geometric framework offers a new avenue for exploring how contact surfaces evolve under asymmetry and external loading.

Technical Summary: Medial Model of Elastic Curve Contacts, from Clasps to Knits

More like this