[Paper Review] Motion by Curvature of Planar Networks

Carlo Mantegazza, Matteo Novaga+1 more|ArXiv.org|Feb 13, 2003

Dynamics and Control of Mechanical SystemsEngineering21 citations

TL;DR

This paper studies the motion by curvature of planar networks, focusing on triods—three smooth curves meeting at a 120° triple junction—with fixed endpoints. It establishes small-time existence, uniqueness, and global regularity of the flow under the Herring condition, proving that embedded triods in strictly convex domains converge smoothly to the Steiner minimal network, avoiding Type I and Type II singularities under uniform length bounds.

ABSTRACT

We consider the motion by curvature of a network of smooth curves with multiple junctions in the plane, that is, the geometric gradient flow associated to the length functional. Such a flow represents the evolution of a two--dimensional multiphase system where the energy is simply the sum of the lengths of the interfaces, in particular it is a possible model for the growth of grain boundaries. Moreover, the motion of these networks of curves is the simplest example of curvature flow for sets which are ``essentially'' non regular. As a first step, in this paper we study in detail the case of three curves in the plane concurring at a single triple junction and with the other ends fixed. We show some results about the existence, uniqueness and, in particular, the global regularity of the flow, following the line of analysis carried on in the last years for the evolution by mean curvature of smooth curves and hypersurfaces.

Motivation & Objective

• To establish existence, uniqueness, and global regularity of the motion by curvature for planar networks with triple junctions.
• To model two-dimensional multiphase systems where energy is the total interface length, such as grain boundary evolution.
• To extend the theory of curvature flow to non-smooth, singular sets, particularly networks with junctions.
• To investigate the asymptotic behavior of such flows, especially convergence to minimal configurations.
• To address open problems in Brakke flow theory for networks, including singularity formation and topological changes.

Proposed method

• Uses Brakke's definition of curvature flow to ensure dimensionality and avoid fattening, while maintaining geometric consistency.
• Applies parabolic PDE theory and a priori estimates to prove short-time existence of smooth solutions for triods with 120° junctions.
• Employs geometric evolution equations for curvature and normal velocity, with the flow governed by $\partial_t γ = k \nu$.
• Implements blow-up analysis and classification of self-similar solutions (homothetic and translating) to study potential singularities.
• Uses the function $E$ and area estimates to rule out Type II singularities under curvature and length bounds.
• Applies $L^2$-norm control of curvature via $\int_0^ au \int_{\mathbb{T}_t} k^2 \, ds \, dt < \infty$ to deduce convergence to zero curvature.

Experimental results

Research questions

• RQ1 Can a smooth motion by curvature be established for a triod with three curves meeting at 120° angles at a triple junction, with fixed endpoints?
• RQ2 Under what conditions does the flow remain regular for all time, avoiding singularities?
• RQ3 Can Type II singularities be ruled out for embedded triods in strictly convex domains with uniformly bounded lengths?
• RQ4 Does the flow converge to the Steiner minimal network connecting the three fixed endpoints?
• RQ5 What is the role of the Herring condition (120° junctions) in ensuring uniqueness and regularity of the Brakke flow?

Key findings

• Small-time existence and uniqueness of smooth solutions are established for triods satisfying the Herring condition (120° junctions).
• Global regularity is proven under the assumption that the lengths of the three curves remain uniformly bounded away from zero.
• Type I singularities are ruled out via curvature and length estimates, implying the flow exists for all time.
• Type II singularities are excluded under the assumption that the limit flow is translating or homothetic, based on geometric and integral estimates.
• The flow converges in $C^\infty$ topology to the Steiner minimal network (zero curvature limit) as $t \to \infty$, due to $\int_0^\infty \int_{\mathbb{T}_t} k^2 \, ds \, dt < \infty$.
• The asymptotic limit is the unique minimal connection between the three fixed endpoints, confirming long-term convergence to a critical configuration.