Modeling Particle Transport In Biomedical Flows Using Implicit Geometry Representations.

Area of Science:

• Computational fluid dynamics
• Biomedical engineering
• Scientific computing

Background:

• Particle-based Lagrangian methods are crucial for modeling physiological and biomedical systems, offering insights into fluid dynamics.
• Current methods face significant computational costs, particularly in resolving particle-wall contact in complex anatomical geometries.
• Efficient geometry representation is vital for high-fidelity patient-specific or device-specific computational models.

Purpose of the Study:

• To develop an efficient particle dynamics model for resolving particle-wall contact in complex anatomical structures.
• To reduce the computational expense associated with particle simulations in biomedical applications.
• To maintain high accuracy and geometric fidelity while improving simulation speed.

Main Methods:

• Developed a particle dynamics model utilizing a signed distance field for implicit representation of anatomical features.
• Transformed Lagrangian contact detection into an equivalent Eulerian operation.
• Validated the approach using simplified geometries and a representative simulation of embolic particles in a human vascular segment.

Main Results:

• The signed distance field approach significantly speeds up bulk particle dynamics computations.
• The method efficiently resolves particle-wall contact without substantial impact on accuracy or geometric fidelity.
• Comparison against classical mesh-based contact detection showed improved performance.

Conclusions:

• The developed signed distance field-based model offers a computationally efficient and accurate solution for particle dynamics simulations in biomedical applications.
• This approach enhances the feasibility of high-fidelity modeling for diseases like stroke and for targeted drug delivery systems.
• The transformation of contact detection from Lagrangian to Eulerian operations represents a significant advancement in computational efficiency.