A unified SPH framework for shell-related interactions

This paper proposes a unified Smoothed Particle Hydrodynamics (SPH) framework that utilizes a novel imaginary particle projection strategy and a fluid-analogous contact model to efficiently and accurately simulate diverse interactions involving thin shells, including fluid-shell, solid-shell, and shell-shell contacts.

The Gist

Imagine you are trying to simulate a chaotic scene on a computer: a truck crashing into a half-full oil tanker, causing the metal tank to crumple while the oil inside sloshes violently. Or perhaps a flexible rubber gate bending under the force of a rushing river.

To do this, scientists use a method called SPH (Smoothed Particle Hydrodynamics). Think of SPH as a digital crowd. Instead of drawing a solid wall or a smooth sheet of water, the computer breaks everything down into millions of tiny, invisible marbles (particles). These marbles bounce off each other, push each other, and flow around obstacles.

The Problem: The "Thin Sheet" Dilemma
In the real world, some things are thick (like a brick wall), but others are very thin (like a sheet of metal, a leaf, or a rubber membrane).

• The Old Way: To simulate a thin sheet using the "marble" method, scientists had to build it out of three or four layers of marbles stacked on top of each other to give it fake thickness. This was like trying to build a piece of paper out of three layers of cardboard. It worked, but it was heavy, slow, and computationally expensive.
• The Challenge: When a thin sheet interacts with water or other objects, the math gets messy. If the sheet is just one layer of marbles, the water particles on the other side "see" a hole. The simulation breaks because the water thinks it can pass right through the paper-thin wall.

The Solution: The "Ghost" Trick
The authors of this paper came up with a clever, unified framework to solve this. They introduced a concept they call "Imaginary Shell Contact Particles."

Here is the analogy:
Imagine you are standing on a very thin, one-person-wide bridge (the shell). You are trying to talk to a friend on the other side (the fluid). Because the bridge is so thin, your voice (the mathematical data) might get lost or leak through.

The authors' solution is to have you hold up a magic mirror (the imaginary particle) that projects your image slightly behind the bridge, into the space where your friend is standing.

• How it works: The computer takes a real particle on the thin shell and instantly creates a "ghost" twin just behind it, along the direction the shell is facing.
• The Result: To the fluid particles, the shell no longer looks like a thin, leaky line. It looks like a solid, thick wall because the "ghost" particles fill in the gaps. The water thinks it's hitting a full 3D object, so it bounces off correctly. Meanwhile, the shell itself remains thin and light, saving the computer a massive amount of work.

Handling the Bumps and Crashes
The paper also tackles what happens when these thin shells crash into other objects (like a tank hitting a truck) or when a shell folds onto itself (like a crumpling piece of foil).

They developed a new way to calculate these crashes by treating the collision like a fluid.

• The Analogy: Imagine two people running into each other. Instead of just saying "Stop!" (which causes glitches in computers), they pretend the space between them is filled with a super-thick, invisible jelly. As they get closer, the "jelly" gets denser and pushes them apart gently but firmly. This prevents the objects from passing through each other (interpenetration) while allowing them to slide and bounce naturally.

Why This Matters
The authors tested their new "Ghost Particle" and "Jelly Collision" system on several difficult scenarios:

1. A dam breaking: Water rushing against a flexible rubber gate.
2. A floating plate: Water pushing down on a thin metal sheet.
3. A truck crash: A truck hitting a half-full oil tanker.

In all these cases, their method was faster (because it didn't need extra layers of particles) and more accurate (because the water didn't leak through the thin walls).

In a Nutshell
This paper gives engineers a new, super-efficient toolkit to simulate thin, flexible things interacting with fluids and solids. By using "ghost" particles to fake thickness and a "jelly" logic for collisions, they can now model complex disasters—like oil tank crashes or flood damage—with high precision and without needing a supercomputer to run for a week. It's like upgrading from a heavy, clunky cardboard model to a sleek, high-tech hologram that behaves exactly like the real thing.

Technical Summary

1. Problem Statement

Smoothed Particle Hydrodynamics (SPH) is a powerful mesh-free method for multi-physics simulations, particularly Fluid-Structure Interaction (FSI). However, modeling thin-shell structures (e.g., tanks, gates, plates) within SPH presents specific challenges compared to full-dimensional solids:

• Dimensionality Mismatch: Shells are typically modeled as reduced-dimensional, single-layer particle discretizations, whereas standard SPH fluid-solid coupling assumes full 3D volumes.
• Kernel Truncation: When fluid particles interact with a thin shell, the kernel support domain is often truncated because the shell lacks thickness. This leads to inaccurate density estimation, pressure fluctuations, and particle penetration.
• Contact Complexity: Existing SPH contact models are often designed for solid-solid interactions. Extending these to handle shell-shell, solid-shell, and shell-self contacts (where a shell folds onto itself) is non-trivial due to the lack of volume in the shell representation.
• Lack of Unified Framework: Previous studies addressed specific cases (e.g., two-sided FSI or simple contact) but lacked a unified approach capable of handling one-sided fluid-shell coupling alongside complex contact scenarios simultaneously.

2. Methodology

The authors propose a unified SPH framework that bridges the gap between reduced-dimensional shell models and full-dimensional fluid/solid interactions. The core methodology consists of three main components:

A. Imaginary Shell Contact Particles (Projection Strategy)

To resolve the kernel truncation issue in fluid-shell interactions, the authors introduce a novel concept of imaginary shell contact particles.

• Mechanism: Real shell particles are projected along their local normal direction within the cutoff radius of a fluid particle.
• Volume Definition: The volume (or area in 2D) of these imaginary particles is not arbitrary; it is calculated based on the local shell curvature.
  • In 2D, the area is derived from the intercepted arc length defined by the curvature center.
  • In 3D, the area is derived from the spherical surface patch, utilizing both principal curvatures.
• Effect: This projection maps the reduced-dimensional shell into an effective full-dimensional representation. It preserves kernel completeness for fluid particles near the boundary, ensuring accurate density initialization and pressure transmission without altering the underlying Fluid-Structure Interaction (FSI) dynamics. Consequently, standard fluid-solid coupling algorithms can be used directly.

B. Unified Contact Model (Particle-to-Particle)

The paper develops a contact model for solid-solid, solid-shell, and shell-shell interactions by drawing an analogy to fluid dynamics.

• Contact Density: A "contact density" ($\rho^c$) is computed for particles in contact using a fluid-style density initialization scheme.
• Contact Forces: The resulting contact forces follow a momentum-equation-inspired formulation, similar to the pressure force calculation in fluids.
• Extension to Shells: By combining the contact density formulation with the projection strategy (imaginary particles), the model efficiently handles shell-related contacts. This allows the framework to naturally manage shell-self contact (self-intersection) and complex multi-body interactions without needing separate, ad-hoc contact algorithms.

C. Numerical Implementation

• Governing Equations: The framework uses the Uflyand–Mindlin plate theory for shell dynamics (accounting for transverse shear) and standard Lagrangian formulations for fluids and solids.
• Time Integration: A dual-criteria time-stepping method is employed. Solid dynamics are sub-cycled within fluid acoustic time steps to handle the stiffness difference between fluids and solids.
• Density Re-initialization: A periodic density re-initialization scheme is used to prevent error accumulation over long simulations.

3. Key Contributions

1. Unified Framework: The first SPH framework to simultaneously handle one-sided fluid-shell coupling, solid-shell, shell-shell, and shell-self contact within a single computational model.
2. Imaginary Particle Projection: A novel geometric projection method that converts thin-shell models into full-dimensional equivalents for interaction purposes, solving the kernel truncation problem via curvature-based volume assignment.
3. Analogous Contact Formulation: A robust contact model that treats contact forces similarly to fluid pressure forces, enabling seamless integration with the shell projection strategy.
4. Versatility: The method eliminates the need for distinct algorithms for different interaction types, streamlining complex multi-physics simulations.

4. Results and Validation

The framework was validated through a series of benchmark tests and a complex industrial case study:

• Hydrostatic FSI (Water on Elastic Plate): The method accurately predicted the static deflection of an aluminum plate under water pressure, matching analytical solutions and demonstrating stable pressure transmission without oscillations.
• Dam-Break through Elastic Gate: Simulated a water column impacting a rubber gate. The shell-based model (single layer) achieved accuracy comparable to multi-layer solid models but with significantly reduced computational cost. Results matched experimental data for gate displacement and flow patterns.
• Dam-Break Impact on Elastic Plate (3D): Validated against experimental data for violent free-surface flow impacting a rubber plate. The framework captured complex phenomena like run-up, splashing, and "S-shaped" structural deformation with high fidelity.
• Flow Around Cylinder: Simulated laminar flow ($Re=100$) past a rigid shell cylinder. The results for drag, lift, and Strouhal number showed excellent agreement with established benchmarks and recent SPH studies, proving the method's ability to capture unsteady wake dynamics.
• Block Sliding: Validated the contact algorithm between an elastic solid and a shell boundary. The simulation perfectly matched the theoretical trajectory of a sliding block, confirming no unphysical penetration.
• Three Rings Contact: Demonstrated the capability to handle complex shell-shell and shell-self contacts. The model successfully simulated large deformations and self-intersection of three interacting rings.
• Oil Tank Collision (Industrial Case): A truck collided with a half-filled oil tank car. The simulation successfully captured the coupled dynamics of large structural deformation, fluid sloshing, and contact events, with smooth stress and pressure fields.

5. Significance

This work represents a significant advancement in computational mechanics for several reasons:

• Efficiency: By using a single-layer shell discretization instead of thick solid models, the method drastically reduces the number of particles required, lowering computational costs for thin-walled structures.
• Robustness: The projection strategy ensures numerical stability in fluid-shell coupling, a known difficulty in SPH, by mathematically restoring kernel completeness.
• Industrial Applicability: The successful simulation of the oil tank collision demonstrates the framework's readiness for complex, real-world engineering problems involving coupled FSI and contact dynamics.
• Open Science: The authors have made the computational code publicly available via the SPHinXsys project, facilitating reproducibility and future research in SPH-based shell interactions.

In conclusion, the proposed unified framework provides a stable, accurate, and efficient solution for simulating a wide range of shell-related interactions, bridging the gap between theoretical shell mechanics and practical SPH implementation.