Files in this item

FilesDescriptionFormat

application/pdf

application/pdfSCHUH-DISSERTATION-2018.pdf (15MB)
(no description provided)PDF

Description

Title:Co-design of surface textures and non-newtonian fluids for decreased friction
Author(s):Schuh, Jonathon Kenneth
Director of Research:Ewoldt, Randy H.
Doctoral Committee Chair(s):Ewoldt, Randy H.
Doctoral Committee Member(s):Allison, James T.; Freund, Jonathan B.; Schroeder, Charles M.
Department / Program:Mechanical Sci & Engineering
Discipline:Mechanical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Full Film Lubrication, Reynolds equation, Criminale-Ericksen-Filbey, Surface Textures, Reynolds number
Abstract:Engineering design is typically limited to linear elastic solids or Newtonian fluids; however, design with more complex materials (for example non-linear viscoelastic materials) can provide ways to enhance performance. The main problems for design with non-linear viscoelastic materials are 1)~there is no universal rheological model, 2)~the non-linear behavior depends on model specific parameters or material specific parameters (eg. polymer molecular weight), and 3)~the governing equations are computationally expensive to solve. This can significantly limit the design space for optimization; multiple material classes exhibit non-linear rheological behavior, and selecting a given class should be dictated by the optimization, not ease of computation. The objective of this work is to provide methods for design with rheologically complex materials that connect system and material level design. The main goal is to validate mathematical models for design with non-linear rheology against experiments with polymer solutions. Our approach is to develop design appropriate mathematical models that are material independent at the system level, but can be made material specific at the material level design stage. We will examine our proposed models in the context of lubricated sliding for co-design of surface textures and rheologically complex fluids to enhance friction reduction; however, the design methods presented are general, and can be applied to any design problem where non-linear rheological behavior is important. For the case study of lubricated sliding, we are interested in simultaneously designing the rheological behavior of the fluid and the surface over which it flows. Previous work has show that for Newtonian fluids, textured surfaces (dimples) decrease friction in lubricated sliding contact. We first perform experiments with simple parameterized textured surfaces and a polymer solution to build intuition for the design problem. Surprisingly, and contrary to results with textured surfaces and Newtonian fluids, the normal forces produced are \textit{always} positive, independent of the direction of motion. They are also \textit{not} a simple superposition of hydrodynamics and shear normal stresses. We show that symmetry must be broken to produce normal forces above the shear normal stress effects, and there appears to be an optimal texture for decreasing friction with textures and polymer solutions. We use the experimental results to guide model selection for design optimization. We choose to represent the fluid using the Criminale-Ericksen-Filbey model, and derive (in the thin film limit) a modified Reynolds equation (that includes shear thinning, normal stress, and inertial effects, and which we refer to as the CEF-Reynolds equation) that serves as a lower bound estimate to the experimental results. We use the CEF-Reynolds equation in material independent (system level) design optimization to again build intuition for the rheological behavior needed for reducing friction, and how this changes the optimal texture surfaces. The optimal solutions are obtained using a multi-objective adaptive surrogate modeling optimization method. The multi-objective optimization produces many optimal designs (Pareto set) due to trade offs in minimizing one objective which result in an increase in the other. The Pareto set contain results with and without polymer additive, suggesting that the results and the design problem are non-trivial. Finally, we propose a method for connecting targeted rheological behavior to material level design, where we focus on polymer solutions. We choose to parameterize the rheological material functions using the FENE-P model, which connects the fluid behavior to the polymer-solvent pair, polymer concentration, and polymer finite extensibility. We obtain the FENE-P model parameters from either fitting or molecular properties, and show that either method could be used for material dependent design; however, care must be taken in the design formulation to ensure that the polymer concentration stays in the dilute limit, where the methods used for connecting the FENE-P parameters to material properties is valid. Outside the scope of this work, these validated methods could to be used to manufacture the optimal textures and fluids to test the predictive capabilities of the models.
Issue Date:2018-04-16
Type:Text
URI:http://hdl.handle.net/2142/100988
Rights Information:Copyright 2018 Jonathon Schuh
Date Available in IDEALS:2018-09-04
Date Deposited:2018-05


This item appears in the following Collection(s)

Item Statistics