MEAM Seminar Series Spring 2017
For Fall 2016 Seminars, click here.
Seminars are held on Tuesday mornings beginning at 10:45 am in Wu and Chen Auditorium, in Levine Hall (unless otherwise noted).
To be added to the MEAM Events mailing list (which sends notifications regarding all departmental seminars and events) please email us at email@example.com.
Xiang I. A. Yang, Postdoctoral Researcher, Stanford University
"The Fluctuating Wall Stress in Wall-Bounded Turbulence"
Measuring, simulating and modeling of the fluctuating wall shear stress at Reynolds numbers of practical relevance are of interest to many fields including sand storm modeling, LES (large-eddy simulation) wall modeling, SWBLI (shock wave boundary layer interaction) modeling etc. In order to take direct measurement of the wall stress, the wall-turbulence community is making a collective effort, building a long pipe at Bologna. We, on the other hand, focus on numerical simulating and theoretical modeling of the fluctuating wall stress. Because DNS limits the Reynolds number, to investigate the behavior of the fluctuating wall stress at high Reynolds numbers, we use LES. With the integral wall model modeling the near wall turbulence, the probability distribution function of the wall stress from the wall-modeled large-eddy simulation agrees well with that in a filtered DNS. The sub-grid wall stress, resulting from the unresolved near wall fluid motions, is then to be modeled. Admitting the fact that the wall stress is the end result of the momentum cascading process, modeling the stress that results from the small-scale near wall turbulence requires a model for the momentum cascade. In this context, the hierarchical random additive model is relevant. Assumptions and predictions of this analytical model are discussed. We then present empirical evidence of the scaling laws that are permitted by this model. Those studies focus on canonical boundary layer flows with no roughness nor wall heat transfer. As real-world engineering often has to tackle rough walls, last, we present a systematic LES study of cuboidal roughness roughened turbulent boundary layers and an analytical rough wall model for the flow sheltering in the roughness layer.
Xiang Yang is a postdoctoral researcher at Center for Turbulence Research, Stanford. Yang received his Ph.D. in mechanical engineering in 2016 from the Johns Hopkins University under the mentorship of Dr. Meneveau and Dr. Mittal. His doctoral work focuses on LES wall modeling, rough wall modeling and theoretical modeling of turbulent boundary layers. His current work at CTR is on shock-wave/transitional-boundary-layer interaction.
Nathan Ip, Ph.D. Candidate
Advisor: Kevin Turner
"Experimental Investigation of Polymer Adhesion Mechanics Using a Blister Contact Test"
3:00 p.m., Room 216, Moore Building
The adhesion of thin layers of soft polymers is important in many
applications, such as tapes, microtransfer printing, and bioinspired
adhesives. Traditional adhesion tests based on probe contacts are not
suitable for characterizing thin layers and common separation-based
specimens, such as the peel test, have well-known limitations. The
blister contact test (BCT) was developed in this dissertation to
overcome the limitations of current methods and was used to investigate
the adhesion and separation of several technologically relevant adhesive
systems. In the BCT, a thin sheet was elastically deformed into
adhesive contact with a reference substrate and the contact area was
optically imaged. Modulated pressure was applied to generate both
advancing and receding adhesive contact. Digital image correlation was
used to measure the displacements of the specimen. The strain energy
release rate at the interface was determined from the measured contact
radius, applied pressure, system geometry, and elastic properties of the
specimen using a mechanics model. An analytical mechanics model based
on von Kármán plate theory was developed and used for analysis of the
BCT data. Finite element analysis was used to validate and identify the
range of applicability of the analytical model.
The BCT was used to investigate the adhesion and separation behaviors of three different polymer adhesive systems. First, experiments between a silicone elastomer (polydimethylsiloxane – PDMS) and a stiff substrate were performed to investigate rate effects in adhesion and separation. For the first time, the rate dependence during advancing contact was characterized. Second, the effect of acid-base interactions on performance of pressure sensitive adhesives (PSAs) was examined via a series of BCTs in which adhesion between different formulations of adhesives and multiple substrates was investigated. Viscoelastic contributions to PSA adhesion were also studied. Finally, the effect of layer thickness on rate dependence was investigated through experiments between polyethylene terephthalate (PET) sheets and PDMS films of different thicknesses. The work in this dissertation demonstrates the flexibility and capability of the BCT as a method to characterize adhesion of flat polymer sheets and provides new understanding of several types of polymer adhesive contacts.
Celia Reina, William K. Gemmill Term Assistant Professor, Department of Mechanical Engineering and Applied Mechanics, University of Pennsylvania
"Multiscale Modeling and Simulation: Some Challenges and New Perspectives"
The design and optimization of the next generation of materials and applications strongly hinge on our understanding of the processing-microstructure-performance relations; and these, in turn, result from the collective behavior of materials’ features at multiple length and time scales. Although the modeling and simulation techniques are now well-developed at each individual scale (quantum, atomistic, mesoscale and continuum), there remain long-recognized grand challenges that limit the quantitative and predictive capability of multiscale modeling and simulation tools. In this talk we will discuss three of these challenges and provide solution strategies in the context of specific applications. These comprise (i) the homogenization of the mechanical response of materials in the absence of a complete separation of length and/or time scales, for the simulation of metamaterials with exotic dynamic properties; (ii) the collective behavior of materials’ defects, for the understanding of the kinematics of large elasto-plastic deformations; and (iii) the upscaling of non-equilibrium material behavior for the modeling of phase change materials.
Celia Reina is the William K. Gemmill Term Assistant Professor in Mechanical Engineering and Applied Mechanics at the University of Pennsylvania. She joined in 2014 after holding the Lawrence Postdoctoral Fellowship at Lawrence Livermore National Laboratory and the HCM Postdoctoral Fellowship at the Hausdorff Center of Mathematics in Bonn, Germany. Dr. Reina received her PhD from the California Institute of Technology in Aeronautics in 2011, with Prof. Michael Ortiz, following a B.S. in Mechanical Engineering from the University of Seville in Spain, and a Master in Structural Dynamics from Ecole Centrale Paris in France.
George I. Park, Engineering Research Associate, Center for Turbulence Research, Stanford University
“Wall Modeling in Large-Eddy Simulation”
Turbulence is the rule, not the exception, in many complex engineering systems. Accurate prediction of turbulent flows over complex wall geometries allows engineers to design fuel-efficient aircrafts, quieter drones, and less expensive, more productive turbomachineries and wind farms. Despite the fact that we know the exact governing equations of fluid motion, and that the techniques to solve them numerically in complex geometries have matured enough over the past decades, first-principle-based prediction methods (such as large-eddy simulation (LES)) have long been in disfavor for solving industry-strength problems due to the cost consideration.
The purpose of this talk is to introduce such cost-related issues in high-fidelity numerical simulation of practical turbulent flows, and then to describe the speaker’s coordinated activities toward predictive and affordable LES in geometrically flexible computational framework. Wall modeling does not attempt to resolve the computationally demanding near-wall region of turbulent flows directly, but instead it provides an alternative boundary closure that can augment the near-wall turbulence, represented with very coarse grids, to the correct state. In this talk, I will convey and expose the basic concepts and methodologies of wall modeling to the general audience. I will first introduce the basic philosophy of wall modeling in LES, and walk the audience through the state-of-the-art wall modeling techniques, showing examples of its application ranging from simple academic flows to an industry-strength problem involving a full 3-D aircraft. Challenges related to computational science such as load balancing, coupling of two parallel unstructured-grid solvers, and its verification will be also discussed briefly. Time permitting, I will discuss an extension of my research to an on-going multi-physics project, where WMLES is to be coupled with particle transport and thermal radiation for predicting the efficiency of a particle-based solar-energy absorption device.
George Park is an engineering research associate in the Center for Turbulence Research (CTR) at Stanford University. Dr. Park received his PhD in Mechanical Engineering from the same institution in 2014 with Prof. Parviz Moin, specializing in unstructured-grid simulation of complex wall-bounded turbulent flows on massively parallel supercomputers. Dr. Park subsequently conducted a postdoctoral research at CTR in modeling of subgrid-scale transport of inertial particles in turbulent flows. Prior to his studies at Stanford, he earned an undergraduate degree (B.S. in Mechanical Engineering,) at Seoul National University in Korea. His current research activities include LES of a full-aircraft flow, particle-laden turbulent flows, non-equilibrium turbulent boundary layer, and surfactant transport in two-phase flows.
Paris Perdikaris, Post-doctoral Associate, Department of Mechanical Engineering, Massachusetts Institute of Technology
“Data-driven Probabilistic Modeling and High-Performance Computing: Algorithms and Applications to Physical and Biological Systems”
The analysis of complex physical and biological systems necessitates the accurate resolution of interactions across multiple spatio-temporal scales, the consistent propagation of information between concurrently coupled multi-physics processes, and the effective quantification of model error and parametric uncertainty. Addressing these grand challenges is a multi-faceted problem that poses the need for a highly sophisticated arsenal of tools in stochastic modeling, high-performance scientific computing, and probabilistic machine learning. Through the lens of three realistic large-scale applications, this talk aims to demonstrate how the compositional synthesis of such tools is introducing a new paradigm in scientific discovery. First, we present multi-scale blood flow simulations in the human brain, and show how high-order methods, massively parallel computing, and concurrent coupling of multi-physics solvers can uncover intrinsic physiological mechanisms in health and disease. We will demonstrate how the introduction of probabilistic machine learning techniques, and the key concept of multi-fidelity modeling, provide a scalable platform for information fusion and lead to significant computational expediency gains. The second application involves an environmental study that illustrates how machine learning tools enable the synergistic combination of simulations, noisy measurements and empirical models towards quantifying the anthropogenic effect of the increasing acidification of coastal waters, and developing a cost-effective monitoring and prediction mechanism. Lastly, we consider the shape optimization of super-cavitating hydrofoils of an ultrafast marine vessel for special naval operations. Specifically, we show how the combination of turbulent multi-phase flow simulations and the concept of multi-fidelity Bayesian optimization allows us to tackle complex engineering design problems in which a rigorous assessment of uncertainty and risk becomes critical in policy and decision making.
Paris Perdikaris received his PhD in Applied Mathematics from Brown University in May 2015. His expertise lies in probabilistic machine learning, computational fluid dynamics, multi-fidelity modeling, uncertainty quantification, and parallel scientific computing. While at Brown he developed scalable machine learning algorithms for predictive multi-fidelity modeling of high-dimensional systems. A parallel research thrust involved developing mathematical models for simulating cardiovascular fluid flows and assessing the characteristics of cerebral pathologies such as cerebral aneurysms in-silico. In June 2015, he moved to MIT as a post-doctoral research associate at the department of Mechanical Engineering and the MIT Sea Grant College Program. His research at MIT is focused on designing a scalable data-driven framework for uncertainty quantification, inverse problems, design optimization, and beyond. The developed algorithms are currently used for risk-averse design optimization of super-cavitating hydrofoils, as well as data assimilation of noisy measurements in coastal regions. Moreover, he has co-advised an undergraduate student working on active learning and data acquisition under uncertainty, and a masters student working on deep learning techniques for object recognition, tracking, and autonomous marine navigation. From 2010-present he has been actively involved in several research projects funded by major US agencies including DOE, AFOSR, NIH and DARPA.
February 17: MEAM/PICS SEMINAR
**THIS SEMINAR WAS RESCHEDULED FROM FEBRUARY 10 DUE TO INCLEMENT WEATHER**
Christian Linder, Assistant Professor, Department of Civil & Environmental Engineering, Stanford University
“Beyond inf-sup: Stability Estimates for Multi-field Variational Principles”
2:00 p.m., Room 337, Towne Building
It is well known that mixed finite element methods have to satisfy certain criteria to provide solvability and stability. The latter criterion is, in the classical context of two-field saddle-point problems such as Stokes flow or quasi-incompressible elasticity, ensured by finite element types that satisfy the well-known inf-sup condition to ensure mesh-independent stability estimates. A number of finite element methods for novel multi-physics applications such as coupled Cahn-Hilliard-type flow in elastic media, extended phase-field models for fracture, poroelasticity or topology optimization as well as gradient-extended plasticity models have a similar saddle-point structure. However, they correspond to a multi-field variational principle and only some of them suffer from similar instabilities. The question as to whether stability estimates are satisfied in these cases for standard discretizations and, if not, how conditions can be obtained that satisfy these estimates, will be discussed in this presentation. Several multi-physics problems developed in our group, that possess a similar saddle-point structure, are investigated with respect to this proposed method. For these examples the satisfaction of the corresponding conditions is shown by means of numerical tests and novel element types for poroelasticity, recently proposed by our group, that are based on incompatible modes and subdivision methods.
Professor Christian Linder is the principal investigator of the Computational Micromaterials Lab at Stanford University. He received his Ph.D. in Civil and Environmental Engineering from UC Berkeley, an MA in Mathematics from UC Berkeley, an M.Sc. in Computational Mechanics from the University of Stuttgart, and a Dipl.-Ing. degree in Civil Engineering from TU Graz. Before joining Stanford in 2013 he was a Junior-Professor of Micromechanics of Materials at the Applied Mechanics Institute of Stuttgart University where he also obtained his Habilitation in Mechanics. Notable honors include a Fulbright scholarship, the 2013 Richard-von-Mises Prize, the 2016 ICCM International Computational Method Young Investigator Award, and the 2016 NSF CAREER Award.
Abtin Rahimian, Postdoctoral Research Associate, Courant Institute of Mathematics, New York University
"Fast Algorithms for Fluid-Structure Interation in Stokes Flow"
The mechanics of complex materials are typically characterized by interacting physical processes, dynamic boundaries, and close coupling over a wide span of spatial and temporal scales. Predictive computational models of such material inherit these characteristics and require many novel algorithmic components. In this talk, I will identify some common features and challenges in predictive modeling of processes with dynamic boundaries in the context of complex fluids, specifically, particulate Stokes flow and intracellular hydrodynamics. These models are instrumental in better understanding the dynamics of bubbles, droplets, particulate hemodynamics, and vesicles as well as the dynamics of intracellular processes. I will discuss the challenges in the numerical simulation of the behavior of such complex materials and will present a family of algorithms that address these challenges for certain class of problems. In particular, I will discuss particulate flows and fibrous flows.
This is a joint work with Denis Zorin, Michael Shelley, and George Biros.
Abtin Rahimian is a Postdoctoral Research Associate at the Courant Institute of Mathematical Sciences at New York University. He earned his Ph.D. in Computational Science and Engineering from Georgia Institute of Technology under the supervision of Professor George Biros. His thesis work primarily focused on designing parallel algorithms for direct numerical simulation of cellular-scale hemodynamics using boundary integral methods. He holds M.Sc. in Mechanical Engineering from University of Pennsylvania and M.Sc. in Mathematics from Georgia Institute of Technology. After receiving his Ph.D., he worked at Goldman Sachs and WorldQuant.
His research interests include scientific computing, multi-scale computational modeling, cellular biomechanics, and parallel algorithms for physical simulations. His current research projects include Tensor-Train accelerated solvers for structured matrices, large-scale boundary integral solvers for partial differential equations in complex and moving geometries, microstructure optimization and design, and modeling mesoscale biophysical systems.
Rahimian is a recipient of the IEEE/ACM SC10 Gordon Bell award from the Association for Computing Machinery’s for “outstanding achievement in high-performance computing applications.”
Helen Minsky, Ph.D. Candidate
Advisor: Kevin Turner
"Composite Posts for Enhanced and Tunable Adhesion"
9:15 a.m., Room 307, Levine Hall
Tunable adhesion is the ability for the same surface to have high adhesion under one set of conditions and low adhesion under another. It has a variety of applications, including transfer printing of micro- and nano-scale components, climbing and perching robots, and material handling in manufacturing. Approaches to tunable adhesion, including the work in this dissertation, often rely on van der Waals forces to achieve dry. Previous strategies for dry tunable adhesives have generally exploited complex fibrillar structures that are inspired by nature. The work in this dissertation investigates a different strategy for enhanced and tunable adhesion based on composite structures with simple geometries.
This dissertation examines the use of composite posts, consisting of stiff insets surrounded by a compliant shell, as an approach for achieving enhanced and tunable adhesion. This composite structure has a high effective adhesion strength under normal loading and low adhesion when shear is applied. Experiments as well as finite element (FE) analysis are used to understand the mechanics of these posts under both types of loading. The adhesion of composite posts is affected by the stress distribution at the contacting surface. Homogeneous posts have concentrated stress near the edge, making crack initiation easy, while the composite posts geometry can result in a redistribution of this stress towards the center, resulting in higher adhesion. The basic mechanics of these posts are demonstrated through experiments on mm-scale posts. The composite mm-scale composite posts have 3x higher adhesion than homogenous posts under normal loading and shear displacement was shown to significantly decrease the effective adhesion strength. Micro-scale posts are studied and used in micro-transfer printing applications. These posts have an effective adhesion strength of 1.5 MPa, and the pull-off force of the composite post is 9x that of a homogeneous post. In both the mm-scale and micro-scale studies, the experimental results are supported by FE simulations. Arrays of micro-scale posts were fabricated and their adhesion behavior characterized. In an array, the contact of each individual post becomes less critical and can contact diverse surfaces. This work established the mechanics of composite posts for achieving enhanced and tunable adhesion.
Michael Posa, Ph.D. Candidate, Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
“Optimization for Control and Planning of Multi-Contact Dynamic Motion”
Whether a robot is assisting a person to move about the home, or packing containers in a warehouse, the fundamental promise of robotics centers on the ability to productively interact with a complex and changing environment in a safe and controlled fashion. However, current robots are largely limited to basic tasks in structured environments--operating slowly and cautiously, afraid of any incidental contact with the outside world. Dynamic interaction, encompassing both legged locomotion and manipulation, poses significant challenges to traditional control and planning techniques. Discontinuities from impact events and dry friction make standard tools poorly suited in scenarios with complex or uncertain contacts between robot and environment. I will present approaches that leverage the interplay between numerical optimization and the mathematical structure of contact dynamics to avoid the combinatorial complexity of mode enumeration. This will include a tractable algorithm for trajectory optimization, without an a priori encoding of the contact sequence, and an approach utilizing sums-of-squares programming to design and provably verify controllers that stabilize systems making and breaking contact.
Michael Posa is a Ph.D. candidate in Electrical Engineering and Computer Science at the Massachusetts Institute of Technology, with an expected graduation in May of 2017. At MIT, he is a member of the Robot Locomotion Group working with Professor Russ Tedrake. He received his B.S. and M.S. in Mechanical Engineering from Stanford University in 2007 and 2008, where he received the Frederick E. Terman Award. Before joining MIT, he worked as an engineer at Vecna Robotics in Cambridge, Massachusetts, designing control systems and simulation tools for the humanoid BEAR robot and other devices. His research emphasizes computational approaches for control and planning of robotic systems with frictional contact. He is a recipient of the Rolf Locher Graduate Fellowship and received the Best Paper award at HSCC in 2013.
Francis D. Lagor, Ph.D. Candidate, Department of Aerospace Engineering, University of Maryland
"Guidance, Estimation, and Control for Robots in Uncertain Fluid Flows"
Some of the largest challenges currently facing mobile robotic platforms involve interactions with surrounding fluid environments. For example, strong gusts can have devastating effects on micro-air vehicles. Autonomous ocean sampling vehicles must efficiently navigate strong, uncertain currents while also conserving energy. Operating in fluid environments is exceptionally difficult due to flow-field nonlinearities, the complexities of fluid mechanical models, and the presence of turbulence. Are there strategies for optimally sensing the surrounding flow to reduce uncertainty in the state of the flow field quickly? How can a vehicle rapidly perform accurate flow-field calculations on-board, given time constraints and hardware limitations? Given flow estimates and uncertainties, what actions can a vehicle take to maximize continued observability of the flow field while maintaining control authority? To address these questions, effective algorithms for control and path planning that incorporate flow-sensing information must be developed. My technical approach is to optimize sensor routing and placement for flow-field observability using reduced-order flow models, perform nonlinear/non-Gaussian estimation of the flow with Bayesian inference, and recursively update flow-field estimates with observer-based feedback control. This approach is effective for long-range path planning as well as vehicle-scale flow sensing and control. I present an example of flow-field estimation in a two-vortex flow using a controlled Lagrangian sensor navigating according to an observability-based guidance strategy with a novel planning measure known as the empirical augmented unobservability index. I will also briefly present other related research interests including autonomous flow sensing and control of a rotorcraft test vehicle near the ground plane, as well as observer-based feedback control of bio-inspired robotic fish executing primitive maneuvers based on flow sensing information.
Frank Lagor is a Ph. D. candidate in the Department of Aerospace engineering and the University of Maryland (graduating May 2017). He received his B.S. in Mechanical Engineering from Villanova University in 2006, and his M.S. in Mechanical Engineering and Applied Mechanics from the University of Pennsylvania in 2009. He worked for Lockheed Martin Space Systems Company for three years prior to pursuing a Ph.D. in dynamics and control. His research interests focus on guidance, estimation, and control of robotic systems in complex flow environments.
April 18: ELSEVIER DISTINGUISHED LECTURE IN MECHANICS
Howard A. Stone, Donald R. Dixon '69 and Elizabeth W. Dixon Professor, Department of Mechanical and Aerospace Engineering, Princeton University
"Seeking Simplicity in the Flows of Complex Fluids"
Fluid mechanics is a discipline with rich phenomena, spanning a wide range of laminar and turbulent flows, instabilities, and applications in industry, nature, and biology and medicine. I will provide examples of our work highlighting (i) new features of classical instabilities triggered by changes in geometry, (ii) multiphase flows relevant to the design of liquid-infused substrates exhibiting effective slip, and, if there is a time, (iii) unexpected dynamics in flow at a T-junction.
Howard A. Stone is the Donald R. Dixon ’69 and Elizabeth W. Dixon Professor in Mechanical and Aerospace Engineering at Princeton University. Stone is a fluid dynamicist who uses experiments, theory and numerical simulations to study transport problems at the intersections of engineering, biology, physics and applied mathematics. Stone received the Bachelor of Science degree in Chemical Engineering from the UC Davis in 1982 and the PhD in Chemical Engineering from Caltech in 1988. In 1989 Stone joined the faculty of the School of Engineering and Applied Sciences at Harvard University, where he eventually became the Vicky Joseph Professor of Engineering and Applied Mathematics. In 2000 he was named a Harvard College Professor for his contributions to undergraduate education. In July 2009 Stone moved to Princeton University. He is a Fellow of the APS and is past Chair of the Division of Fluid Dynamics. In 2008 he was the first recipient of the G.K. Batchelor Prize in Fluid Dynamics and in 2016 he received the APS Fluid Dynamics Prize. He was elected to the National Academy of Engineering in 2009, the American Academy of Arts and Sciences in 2011, and the National Academy of Sciences in 2014.