Discrete Particle Dynamics Models: Computational Aspects And Applications

<p>Particle dynamics considers discrete particles as Newtonian point masses in diverse domains such as, pedestrian particles in pedestrian movement modeling and molecules in material science modeling. These dynamic modeling approaches share a commonality in computational approach where a series of ordinary differential equations are solved for the particle’s motion and equilibrium under active interactions and boundary conditions. In general, particle dynamics approaches are computationally intensive and address problems that require a vast parameter design space. In this dissertation, a computational paradigm is developed for large-scale particle dynamics modeling and demonstrated by three process design applications, (a) infection risk modeling for mitigation policy design, (b) powder spread modeling for metal additive manufacturing design, and (c) energy characterization of nanomaterials modified fiber-reinforced composite for interface design. All these problems require large computational resources to parametrically address inherent uncertainties. Three major strategies are incorporated to address the inherent uncertainties and the resulting computational cost issues. (a) Infection risk modeling incorporates conventional parameter sweeps and empirical data to address the stochastic human behaviors. The incorporation of pedestrian movements suggests the need for more stringent guidelines to reduce transportation-related infectious disease. (b) Powder spread modeling formulates a strategic parametric analysis approach combining a novel algorithmic parameter sweep and artificial intelligence methods to capture the correlation between modeling parameters and outcomes for parameter space interpolation and engineering process design. Such analysis strategy requires at least ten times fewer computational resources than conventional methods while achieving a comprehensive understanding of the parameter space. Modeling outcomes provide efficient and insightful iv guidance for metal additive manufacturing process design. (c) Interface characterization of nanomaterials at fiber-reinforced composite interface utilizes data fusion approaches to address the unknown parameters of particulate systems. A composite interface composing jute fibers and hydroxyapatite nanocrystals is designed to demonstrate the data fusion approach integrating experimental data with modeling calibration. The nanoparticle enhanced interface endows the hybrid composite with superior mechanical performance and functionalizes jute fibers for additive manufacturing. Data assimilation from experiments enable model calibration and validation. The calibrated model allows for comprehensive parametric and sensitivity analysis, which aims to provide reliable insights for future composite interface design. Overall, the computational strategies for discrete particle dynamics models demonstrated in this dissertation are readily expandable for other systems and process design tasks involving vast design space.</p>

Name thesis:degree_name

Doctor of Philosophy in Aerospace Engineering

Level thesis:degree_level

Dissertation - Open Access

Discipline thesis:degree_discipline

Aerospace Engineering

Year

2024

Author dc:creator

• Wu, Yuxuan

• Discrete element method
• Process design
• Parametric analysis
• Complexed system modeling
• Structures and Materials
• Systems Engineering and Multidisciplinary Design Optimization

Repository record dc:identifier

https://commons.erau.edu/edt/855

OAI identifier oai:identifier

oai:commons.erau.edu:edt-1887