Flow manipulation in soft systems: biophysics and microfluidic discovery through scalable software design

Parthasarathy, Tejaswin

Permalink

https://hdl.handle.net/2142/120479 Copy

Description

Title

Flow manipulation in soft systems: biophysics and microfluidic discovery through scalable software design

Author(s)

Parthasarathy, Tejaswin

Issue Date

2022-12-21

Director of Research (if dissertation) or Advisor (if thesis)

Gazzola, Mattia

Doctoral Committee Chair(s)

Gazzola, Mattia

Committee Member(s)

• Chamorro, Leonardo
• Hilgenfeldt, Sascha
• Panesi, Marco

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• fluid dynamics
• fluid--structure interaction
• viscous streaming
• scalable software
• HPC
• elasticity
• structure-dynamics
• resonance
• parallelism
• flexible systems
• Cosserat rods
• elastohydrodynamics
• dynamical systems
• nonlinear dynamics
• inertial microfludics
• flow manipulation
• active-matter
• magnetism
• soft-robotics
• soft systems
• micro-robotics
• compliant

Abstract

Interaction between flow and soft structures pervades biology and engineering across scales, spanning domains from bio-medical/physiological flows to microfluidics, where there exists an inextricable nexus between compliant mechanics, flow environment interaction, control & behaviour. Modeling this flow–structure interaction helps dissect underlying mechanisms to yield insights into rational flow manipulation for informing engineering decisions while designing devices. This has potential applications beyond the above domains: from medicine, where compliant devices may be used for drug delivery, to inertial microfluidics, where streaming effects can be used for flow control applications such as particle manipulation & trapping. In this context, the overarching goal of this thesis is to establish algorithmic and software capabilities for simulating the dynamics of flow, compliant structures and their coupled interaction, and demonstrate its utility in designing rational flow control applications. In particular, I focus on regimes characterized by equally important inertial, viscous, and elastic effects in which soft biological organisms, pliable microfluidic devices, and part-artificial, part-living mini-bots reside. Towards this vision, I organize my research activities into three main thrusts: Rational manipulation of flow environment via viscous streaming. This thrust focuses on flow control with a particular emphasis on viscous streaming. This fluid mechanism takes place when an immersed body oscillates within specific size–frequency ratios, and is responsible for the emergence of characteristic rectified flows via non-linear fluid responses. These structures offer appealing flow control options for particle manipulation, mixing and lysis. The streaming size– frequency ratios overlap with the regimes and applications of our interest, hence our motivation to investigate these flows. Here, I first establish the potency of rigid-body streaming as a robust flow control mechanism in the context of particle transport applications. I demonstrate this by leveraging streaming to enable and enhance delivery of drug payloads in a minimal drug delivery setting by a cylindrical mini-bot, even in the presence of unsteady flow effects such as wakes and shear layers. I then illustrate how (hitherto unexplored) body curvature effects can be exploited in streaming-based applications by designing the bot geometry to favorably sculpt the streaming flow, and in turn payload delivery, in these uncertain environments. This relation between body geometry and streaming flow topology can be systematically uncovered by employing tools from dynamical systems theory, and provides basis for rational manipulation of streaming flow for applications beyond particle transport. I provide an example of such an application, where by collaborating with experimentalists, a streaming-based inertial-particle separator was developed. Dynamics of soft structures immersed in flows. This thrust shifts attention from dynamics of the (exogenous) flow environment to the (endogenous) body mechanics. Indeed, soft and compliant structures are ubiquitous in the systems of our interest across scales, which then necessitates robust, validated, and versatile algorithms to resolve them. Given the paucity of such versatile algorithms in literature, I developed one for solving incompressible flow–structure interaction (FSI) problems for mixed rigid/soft hyperelastic bodies within a consistent framework based on the remeshed-vortex method. Through multiple benchmarks, I demonstrate the accuracy, robustness and versatility of the above monolithic solver (where flow and structure equations are couched as a single set of governing equations over the domain). I then utilize this solver to investigate oscillations and streaming from soft bodies, where I showcase that body elasticity offers another dimension to rationally manipulate, and hence design streaming topology for flow control applications. Nevertheless soft-robotic and biological architectures are not only compliant but are heterogeneous in their elastic constitution. That is, they are usually composed of different components distributed across their body leading to an anisotropic deformation response not readily captured in our monolithic framework. Then, to capture these features, and drawing from the observation that filament-based organization of elements is common in the systems of interest, one can adopt rod theories to capture the 3D dynamics of elastic filaments via inexpensive 1D Lagrangian representations. These filaments can then be organized and connected to synthesize active biological and engineered layouts, which can be leveraged for flow manipulation. Performing such filament simulations necessitates a cohesive software framework and to date, no established reference software exists. An open, performant software environment for flow–structure dynamics. To respond to this need, I kickstarted Elastica, an open-source ready-to-use ecosystem for the simulation, analysis, design and control of dynamic, heterogeneous structures made of slender, elastic rods. As realistic architectures are comprised of many elastic rods each, I first outfit Elastica with a highly performant, scalable simulation backend targeting high-fidelity, tens-to-hundreds gigaflops-grade simulations on a single node entailing up to a million filaments, three-orders more than the current state-of-the-art. I achieve this via high-performance computing (HPC) strategies relying on a new block architecture, implemented via a novel star hierarchy to enable platform-independent, scalable performance in Lagrangian simulation codes such as Elastica. I then demonstrate the utility of Elastica as a flexible, comprehensive and validated simulation toolkit in a range of biological and engineered systems across scales and environments, by performing some of the largest ever flexible system simulations. These demonstrations, spanning fibrous biological passive and active-matter to engineered magnetized microbots, transcend scientific domains from bio-mechanics to robotics and control. Finally, I illustrate a route towards consolidating the modeling flexibility afforded by Lagrangian Elastica rods powered by blocks with the versatility of our monolithic FSI algorithm, thus enabling accelerated simulations of realistic immersed soft architectures for scientific discovery and engineering design.

Graduation Semester

2023-05

Type of Resource

Thesis

Copyright and License Information

© 2022 Tejaswin Parthasarathy

Owning Collections