A computational model for short-range van der Waals interactions between beams and shells

1. Problem Statement

The paper addresses the computational modeling of short-range van der Waals (vdW) interactions between deformable fibers (modeled as beams) and membranes (modeled as shells).

• Context: These interactions are critical in biological systems (e.g., protein-membrane adhesion, DNA folding) and nanotechnology (e.g., electrospun nanofibers, carbon nanotubes on substrates).
• Challenge: Accurately modeling these forces requires integrating a pairwise potential over the volumes of both bodies. For a beam and a shell, this results in a 6-dimensional (6D) integral, which is computationally prohibitive for finite element analysis (FEA), especially given the steep gradients of interaction forces at small separations.
• Gap: Existing literature lacks efficient, accurate computational models for beam-shell contact based on potential fields, often relying on simplified 2D or axisymmetric assumptions, or neglecting the complex geometry of 3D deformable bodies.

2. Methodology

The authors propose a coarse-grained approach that reduces the computational complexity while maintaining accuracy for short-range interactions.

A. Theoretical Framework

• Potential Model: The interaction is modeled using the Lennard-Jones (LJ) potential, which combines attractive vdW forces (inverse 6th power) and steric repulsion (inverse 12th power).
• Structural Models:
  • Beam: Bernoulli-Euler (BE) beam theory (accounts for axial, bending, and torsional deformation).
  • Shell: Kirchhoff-Love (KL) shell theory (accounts for membrane and bending strains).
• Discretization: The system is solved using Isogeometric Analysis (IGA) with quartic B-splines ($C^1$ continuity), which is well-suited for handling the high-order derivatives required by shell and beam theories.

B. The Core Innovation: Surrogate Beam-Shell (SBS) Model

To reduce the 6D integral to a manageable form, the authors introduce a Section-Surrogate Plate concept:

1. Pre-integration: The 6D volume integral is split. The integration over the beam's cross-section and the shell's local volume is performed analytically.
2. Surrogate Plate Approximation: Instead of integrating over the curved shell volume, the shell is locally approximated at the closest point by an infinite flat plate (surrogate plate) tangent to the shell surface.
3. Reduction: This reduces the problem to a 1D numerical integration along the beam's axis. The interaction potential at any point on the beam axis is calculated as the interaction between a disk (beam cross-section) and an infinite plate (surrogate shell).
4. Variational Formulation: The authors derive the weak form of the equilibrium equations, including the variation of the interaction potential with respect to beam displacement, shell displacement, and the orientation angle between the beam cross-section and the surrogate plate.

C. Error Analysis

The paper provides a rigorous error estimate for the "disk-flat plate" approximation:

• Small Separations: The error converges to zero as the gap approaches zero.
• Curvature Effects: The error is bounded and depends on the ratio of the shell radius to the beam radius ($R_{shell}/R_{beam}$) and the shell thickness. For $R_{shell}/R_{beam} > 10$, the relative error drops below 1%.
• Scaling: The SBS approach correctly captures the scaling of interaction forces ($\propto g^{-2}$) in the critical short-range regime, which is essential for adhesion and peeling phenomena.

3. Key Contributions

1. Novel Computational Formulation: Development of the Surrogate Beam-Shell (SBS) model, which reduces the interaction potential integration from 6D to 1D via analytical pre-integration of a disk-plate law.
2. Analytical Derivations:
   • Derivation of the disk-sphere vdW law (Appendix F), providing a reference solution for error estimation.
   • Derivation of the variation and linearization of the interaction potential, including three formulations:
     • Full Formulation (FF): Includes all geometric variations.
     • Reduced Formulation 1 (RF1): Neglects angle variation.
     • Reduced Formulation 2 (RF2): Neglects both angle variation and cross-section orientation (simplest, fastest).
3. Benchmark Examples: Two novel numerical examples are presented to validate the model, as no direct experimental benchmarks exist in the literature for this specific problem:
   • Peeling: Peeling an adhering beam from a shell.
   • Bending: Bending a shell by an adhering beam under applied moments.

4. Results

• Convergence and Efficiency: The IGA discretization with the SBS model demonstrates excellent convergence. The reduction from 6D to 1D integration offers a massive gain in computational efficiency.
• Peeling Simulation:
  • The model successfully captures the non-linear peeling process, including the transition from adhesion to pull-off.
  • Formulation Comparison: The simplest formulation (RF2) yields results nearly identical to the Full Formulation (FF) during the initial adhesion phase. However, RF2 slightly overestimates the pull-off force because it neglects the changing orientation of the beam cross-section at high curvatures.
  • Parameter Sensitivity: The reaction force scales with the interaction strength ($s_i$) following a power law ($F \propto s_i^{0.678}$).
• Bending Simulation:
  • The model captures the complex interaction where the beam adheres to the shell, causing the shell to deform significantly.
  • Force Distribution: The interaction force exhibits steep gradients near the contact points, highlighting the necessity of the coarse-grained approach to resolve these singularities without excessive mesh refinement.
  • Mesh Independence: The results show that the interaction force is highly sensitive to mesh density, while displacement is less so, guiding the selection of appropriate discretization levels.

5. Significance and Future Work

• Scientific Impact: This work bridges the gap between molecular dynamics (too expensive for macroscopic structures) and classical continuum contact mechanics (often ignores vdW forces). It provides a robust tool for simulating fiber-membrane interactions in biology and nanotechnology.
• Practical Application: The model is directly applicable to the design of nanofiber membranes, bio-inspired materials, and the study of cellular processes involving membrane deformation.
• Limitations & Future Directions:
  • The current model assumes frictionless interaction. Future work will introduce a tangential contact potential to model friction and energy dissipation.
  • The approximation of the shell as a flat plate limits accuracy for very large gaps or highly curved shells relative to the beam, though this is acceptable for the short-range adhesion focus of the paper.
  • The interaction force on the shell is approximated as point forces; future improvements may involve a "two half-pass" approach for better force distribution on the shell.

In summary, the paper presents a highly efficient and accurate computational framework for simulating short-range vdW interactions between beams and shells, enabling the study of complex, non-linear adhesion phenomena that were previously computationally intractable.