2016 Symposium Agenda
Introductory Remarks: S. Jack Hu, Vice President for Research, U-M
What’s Happening at MICDE? Krisha Garikipati, Associate Director, MICDE; Professor of Mechanical Engineering and Mathematics
Speaker: Irene Qualters, director of NSF’s Division of Advanced Cyberinfrastructure
Title: Research Cyberinfrastructure as a Strategic Enabler for Science and Engineering
While it may be self-evident that Cyberinfrastructure is increasingly important to researchers, it may be less obvious that science and engineering also have unique needs. This talk will explore trends in research and how NSF CI investment strategies reflect the resulting priorities.
Speaker: C. Alberto Figueroa, U-M Departments of Surgery and Biomedical Engineering
Title: Surgical Planning in Congenital Heart Disease using Computational Techniques: Procedure Optimization and Clinical Validation
Pulmonary arteriovenous malformations (PAVMs) in congenital heart disease patients is a well-documented observation. PAVMs create a low resistance connection between the arteries and veins in the pulmonary circulation, resulting in abnormally high pulmonary flow and poor blood oxygenation, and are associated with significant morbidity. The mechanism of PAVM development remains unknown, however studies indicate the uneven balance of hepatic venous flow (HVF) between the lungs is critical. The absence of HVF to a lung vasculature results in the formation and persistence of unilateral PAVM. Surgical restoration of HVF to the afflicted side is shown to alleviate the PAVM. However, surgical procedures are invasive, complex, and have variable success.Using our in-house image-based blood flow modeling software CRIMSON (www.crimson.software), we performed a computational analysis of the deployment of a stent-graft in a congenital heart disease patient suffering from severe right-side PAVM. The optimal configuration of the stent-graft implantation was assessed by studying different stent-graft protrusion lengths and the resulting HVF distribution. The results of the computational analysis were subsequently used to guide the intervention performed at the UM Mott Children’s Hospital. Post-intervention verification of the computational predictions was obtained, showing excellent qualitative agreement between clinical measurements and computations (see Figure).
This is one of the first times that image-based computational modeling has been used to optimally plan endovascular procedures, and probably the first study that provides validation to the computations with pre- and post-operative clinical data. The tools described here hold great promise to aid in the planning of a number of endovascular as well as surgical procedures in congenital heart disease patients.
Title: Earth System Modeling on Upcoming Exascale Computers
11:00 a.m. — Break
Title: Direct numerical simulations of particulate flows: breaking barriers, enabling discoveries
From blood flow to subsurface flows, particulate flows are ubiquitous. Direct numerical simulations (DNS) of dense, deformable particle suspensions in viscous fluids are extremely challenging yet critically important to bring insights into their macro-scale flow behavior. In this talk, I will present recent advances made by our group in overcoming several computational bottlenecks, especially those arising in the context of dense suspensions confined by complex geometries. Incorporating stable time-marching schemes, fast and spectrally-accurate algorithms, we were able to simulate the hydrodynamics of over 1000 deformable particles flowing through a complex periodic geometry in less than a minute per time-step on a laptop. A wide range of applications from designing microfluidic chips to optimizing drug-carrier shapes will be discussed.
Speaker: Linda Petzold, Professor in the Department of Computer Science and the Department of Mechanical Engineering; Director of the Computational Science and Engineering Graduate Emphasis, University of California Santa Barbara
Title: Stochastic Simulation at Your Service
Stochasticity plays an important role in many biological processes. At the same time, algorithms and software for discrete stochastic simulation have advanced to the point where not only simulation of well-mixed systems, but spatial stochastic simulation on 3D complex geometries, parameter estimation for stochastic systems, and efficient computation of the probability of rare events, can be performed with accuracy and reliability. A few years ago we embarked on a quest to build a unified software environment to enable biologists to easily harness the power of these tools. We envisioned that users might build an ODE model or discrete stochastic model on a laptop, and scale it up to increasing levels of complexity, accessing tools such as those mentioned above, and deploying computing resources from the cloud with the push of a button when they are needed. StochSS: Stochastic Simulation as-a- Service, is available for download at www.stochss.org. As the capabilities of StochSS have grown, so has our vision of the roles that computing can play in the advancement of science.
12:30 p.m. — Lunch and Poster Session
Advanced Research Computing at Michigan: Eric Michielssen, Associate Vice President — Research Computing, U-M
Title: New Methods for Tracking Evolving Interfaces: Voronoi Implicit Interface Methods with Applications to Industrial Foams, Biological Cell Clusters, and Domain Decomposition
Many scientific and engineering problems involve interconnected moving interfaces separating different regions, including dry foams, crystal grain growth, and multi-cellular structures in man-made and biological
materials. Producing consistent and well-posed mathematical models that capture the motion of these interfaces, especially at degeneracies, such as triple points and triple lines where multiple interfaces meet, is challenging.Joint with Robert Saye of UC Berkeley, we have built the Voronoi Implicit Interface Method (VIIM), which is an efficient and robust mathematical and computational methodology for computing the solution to two and three-dimensional multi-interface problems involving complex junctions and topological changes in an evolving general multiphase system. We demonstrate the method on a collection of problems, including geometric coarsening flows under curvature, incompressible
flow coupled to multi-fluid interface problems, and biological cell cluster growth under competing effects of geometry, fluid dynamics, and elasticity.Finally, we compute the dynamics of unstable foams, such as soap bubbles, evolving under the combined effects of gas-fluid interactions, thin-film lamella drainage, and topological bursting.
Speaker: Ann Jeffers, U-M Department of Civil and Environmental Engineering
Title: Multiphysics simulation to advance our understanding of traveling fires
Major accidental fires (e.g., the World Trade Center fires in 2001 and the Windsor Tower fire in 2005) have provided empirical evidence that large-scale fires tend to travel across floors rather than burn uniformly. Preliminary studies of structural systems under non-stationary and non-uniform fires have shown that structural responses such as load reversal, column instability, and the development of large tensile forces in floor beams could lead to premature structural failure under “traveling fire” scenarios. The preliminary work, however, was based on an idealized traveling fire model in which the fire was assumed to move in a 1D or 2D pattern with constant velocity. The work presented here takes a different approach in which the fire is modeled using computational fluid dynamics and the structural response is modeled using finite element analysis. This multiphysics-based approach allows for the realistic simulation of fire spread from object to object and accounts for system-level structural performance under non-uniform fire exposure. This presentation describes the computational approach that is being taken, including the innovative finite elements and algorithms that have been developed at the University of Michigan for coupling a high-fidelity computational fluid dynamics model to a low-resolution (i.e., macro) finite element model. Simulations are under way to gain an understanding of how traveling fires affect structural performance so that structural designs and forensic investigations can be improved.
3:30 p.m. — Break
Title: Rebooting Computational Engineering, How Technology Transitions Transform our Turf
Technology transitions are never ending. Just when one technology appears to dominate the market, another technology emerges to push it out of the way. Computational engineering is buffeted by these transitions. The algorithms, the directions and the applications of computational engineering are driven by technology trends, while computational engineering in turn speeds the technology transitions along the way.This talk explores how technology transitions shape the computational landscape. Three types of transitions are considered. The first is the evolution of computer hardware, from mainframes to PC’s to the cloud, which provides a changing canvas for developing simulation software. The second is the invention of new algorithms such as adaptive mesh refinement and model order reduction that make earlier algorithms obsolete. And the third is the shift in the dominant engineering problem at hand, such as the shift from defense applications to computer applications, and more recently to mobile device applications, over the past decades.