S. Levent Yilmaz

University of Pittsburgh, Pittsburgh, Pennsylvania, United States

Are you S. Levent Yilmaz?

Claim your profile

Publications (23)36.66 Total impact

  • P.H. Pisciuneri · S.L. Yilmaz · P.A. Strakey · P. Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: A review is presented of the evolution of a massively parallel solver for large eddy simulation (LES) of turbulent reacting flows via the filtered density function (FDF). Development of an efficient parallel implementation is particularly challenging due to the hybrid Eulerian/Lagrangian structure of typical FDF simulators. The performance of a novel parallel simulator is assessed at each of the major steps of its development. Subsequent efforts to improve scaling at each of these stages are discussed along with the prospects for further enhancements.
    No preview · Article · Jan 2015
  • [Show abstract] [Hide abstract]
    ABSTRACT: Production of electricity and propulsion systems involve turbulent combustion. Computational modeling of turbulent combustion can improve the efficiency of these processes. However, large tensor datasets are the result of such simulations; these datasets are difficult to visualize and analyze. In this work we present an unsupervised statistical approach for the segmentation, visualization and potentially the tracking of regions of interest in large tensor data. The approach employs a machine learning clustering algorithm to locate and identify areas of interest based on specified parameters such as strain tensor value. Evaluation on two combustion datasets shows this approach can assist in the visual analysis of the combustion tensor field.
    No preview · Article · Jan 2015 · Mathematics and Visualization
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Because patients can remain colonized with vancomycin-resistant enterococci (VRE) for long periods of time, VRE may spread from one health care facility to another. Using the Regional Healthcare Ecosystem Analyst, an agent-based model of patient flow among all Orange County, California, hospitals and communities, we quantified the degree and speed at which changes in VRE colonization prevalence in a hospital may affect prevalence in other Orange County hospitals. A sustained 10% increase in VRE colonization prevalence in any 1 hospital caused a 2.8% (none to 62%) average relative increase in VRE prevalence in all other hospitals. Effects took from 1.5 to >10 years to fully manifest. Larger hospitals tended to have greater affect on other hospitals. When monitoring and controlling VRE, decision makers may want to account for regional effects. Knowing a hospital's connections with other health care facilities via patient sharing can help determine which hospitals to include in a surveillance or control program.
    Full-text · Article · Aug 2013 · American journal of infection control
  • S.L. Yilmaz · N. Ansari · P.H. Pisciuneri · M.B. Nik · C.C. Otis · P. Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: An overview is presented of recent advances in the filtered density function (FDF) modeling and simulation of turbulent combustion. The review is focused on the developments that have facilitated the FDF to be broadly applied in large eddy simulation (LES) of practical flows. These are primarily the development of a new Lagrangian Monte Carlo solver for the FDF, and the implementation of this solver on Eulerian domains portrayed by unstructured grids. With these developments, it is now much easier to apply FDF for predictions of reacting flows in complex geometrical configurations.
    No preview · Article · Jul 2013 · Journal of Applied Fluid Mechanics
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Objective As healthcare systems continue to expand and interconnect with each other through patient sharing, administrators, policy makers, infection control specialists, and other decision makers may have to take account of the entire healthcare ‘ecosystem’ in infection control. Materials and methods We developed a software tool, the Regional Healthcare Ecosystem Analyst (RHEA), that can accept user-inputted data to rapidly create a detailed agent-based simulation model (ABM) of the healthcare ecosystem (ie, all healthcare facilities, their adjoining community, and patient flow among the facilities) of any region to better understand the spread and control of infectious diseases. Results To demonstrate RHEA's capabilities, we fed extensive data from Orange County, California, USA, into RHEA to create an ABM of a healthcare ecosystem and simulate the spread and control of methicillin-resistant Staphylococcus aureus. Various experiments explored the effects of changing different parameters (eg, degree of transmission, length of stay, and bed capacity). Discussion Our model emphasizes how individual healthcare facilities are components of integrated and dynamic networks connected via patient movement and how occurrences in one healthcare facility may affect many other healthcare facilities. Conclusions A decision maker can utilize RHEA to generate a detailed ABM of any healthcare system of interest, which in turn can serve as a virtual laboratory to test different policies and interventions.
    Full-text · Article · Apr 2013 · Journal of the American Medical Informatics Association
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objective. Implementation of contact precautions in nursing homes to prevent methicillin-resistant Staphylococcus aureus (MRSA) transmission could cost time and effort and may have wide-ranging effects throughout multiple health facilities. Computational modeling could forecast the potential effects and guide policy making. Design. Our multihospital computational agent-based model, Regional Healthcare Ecosystem Analyst (RHEA). Setting. All hospitals and nursing homes in Orange County, California. Methods. Our simulation model compared the following 3 contact precaution strategies: (1) no contact precautions applied to any nursing home residents, (2) contact precautions applied to those with clinically apparent MRSA infections, and (3) contact precautions applied to all known MRSA carriers as determined by MRSA screening performed by hospitals. Results. Our model demonstrated that contact precautions for patients with clinically apparent MRSA infections in nursing homes resulted in a median 0.4% (range, 0%-1.6%) relative decrease in MRSA prevalence in nursing homes (with 50% adherence) but had no effect on hospital MRSA prevalence, even 5 years after initiation. Implementation of contact precautions (with 50% adherence) in nursing homes for all known MRSA carriers was associated with a median 14.2% (range, 2.1%-21.8%) relative decrease in MRSA prevalence in nursing homes and a 2.3% decrease (range, 0%-7.1%) in hospitals 1 year after implementation. Benefits accrued over time and increased with increasing compliance. Conclusions. Our modeling study demonstrated the substantial benefits of extending contact precautions in nursing homes from just those residents with clinically apparent infection to all MRSA carriers, which suggests the benefits of hospitals and nursing homes sharing and coordinating information on MRSA surveillance and carriage status.
    No preview · Article · Feb 2013 · Infection Control and Hospital Epidemiology
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background: Hospital infection control strategies and programs may not consider control of methicillin-resistant Staphylococcus aureus (MRSA) in nursing homes in a county. Methods: Using our Regional Healthcare Ecosystem Analyst, we augmented our existing agent-based model of all hospitals in Orange County (OC), California, by adding all nursing homes and then simulated MRSA outbreaks in various health care facilities. Results: The addition of nursing homes substantially changed MRSA transmission dynamics throughout the county. The presence of nursing homes substantially potentiated the effects of hospital outbreaks on other hospitals, leading to an average 46.2% (range, 3.3%-156.1%) relative increase above and beyond the impact when only hospitals are included for an outbreak in OC's largest hospital. An outbreak in the largest hospital affected all other hospitals (average 2.1% relative prevalence increase) and the majority (~90%) of nursing homes (average 3.2% relative increase) after 6 months. An outbreak in the largest nursing home had effects on multiple OC hospitals, increasing MRSA prevalence in directly connected hospitals by an average 0.3% and in hospitals not directly connected through patient transfers by an average 0.1% after 6 months. A nursing home outbreak also had some effect on MRSA prevalence in other nursing homes. Conclusions: Nursing homes, even those not connected by direct patient transfers, may be a vital component of a hospital's infection control strategy. To achieve effective control, a hospital may want to better understand how regional nursing homes and hospitals are connected through both direct and indirect (with intervening stays at home) patient sharing.
    No preview · Article · Jan 2013 · Medical care
  • P.H. Pisciuneri · S.L. Yilmaz · P.A. Strakey · P. Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: A new computational methodology, termed "irregularly portioned Lagrangian Monte Carlo-finite difference" (IPLMCFD), is developed for large eddy simulation (LES) of turbulent combustion via the filtered density function (FDF). This is a hybrid methodology which couples a Monte Carlo FDF simulator with a structured Eulerian finite difference LES solver. The IPLMCFD is scalable to thousands of processors; thus it is suited for simulation of complex reactive flows. The scalability and consistency of the hybrid solver and the realizability and reliability of the generated results are demonstrated via LES of several turbulent flames under both nonpremixed and premixed conditions.
    No preview · Article · Jan 2013 · SIAM Journal on Scientific Computing
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Efforts to control life-threatening infections, such as with methicillin-resistant Staphylococcus aureus (MRSA), can be complicated when patients are transferred from one hospital to another. Using a detailed computer simulation model of all hospitals in Orange County, California, we explored the effects when combinations of hospitals tested all patients at admission for MRSA and adopted procedures to limit transmission among patients who tested positive. Called "contact isolation," these procedures specify precautions for health care workers interacting with an infected patient, such as wearing gloves and gowns. Our simulation demonstrated that each hospital's decision to test for MRSA and implement contact isolation procedures could affect the MRSA prevalence in all other hospitals. Thus, our study makes the case that further cooperation among hospitals-which is already reflected in a few limited collaborative infection control efforts under way-could help individual hospitals achieve better infection control than they could achieve on their own.
    Full-text · Article · Oct 2012 · Health Affairs
  • [Show abstract] [Hide abstract]
    ABSTRACT: Significant progress has been made in the development of subgrid scale (SGS) closures based on a filtered density function (FDF) for large eddy simulations (LES) of turbulent reacting flows. The FDF is the counterpart of the probability density function (PDF) method, which has proven effective in Reynolds-averaged simulations (RAS). However, while systematic progress is being made advancing the FDF models for relatively simple flows and lab-scale flames, the application of these methods in complex geometries and high speed, wall-bounded flows with shocks remains a challenge. The key difficulties are the significant computational cost associated with solving the FDF transport equation and numerically stiff finite-rate chemistry. For LES/FDF methods to make a more significant impact in practical applications a pragmatic approach must be taken that significantly reduces the computational cost while maintaining high modeling fidelity. An example of one such ongoing effort is at the NASA Langley Research Center, where the first generation FDF models, namely the scalar filtered mass density function (SFMDF), are being implemented into VULCAN, a production-quality RAS and LES solver widely used for design of high speed propulsion flowpaths. This effort leverages internal and external collaborations to reduce the overall computational cost of high fidelity simulations in VULCAN by: implementing the high order methods that allow reduction in the total number of computational cells without loss in accuracy; implementing first generation of high fidelity scalar PDF/FDF models applicable to high-speed compressible flows; coupling RAS/PDF and LES/FDF into a hybrid framework to efficiently and accurately model the effects of combustion in the vicinity of the walls; developing efficient Lagrangian particle tracking algorithms to support robust solutions of the FDF equations for high speed flows; and utilizing finite-rate chemistry parametrization, such as flamelet models, to reduce the number of transported reactive species and remove numerical stiffness. This paper briefly introduces the SFMDF model (highlighting key benefits and challenges), and discusses particle tracking for flows with shocks, the hybrid coupled RAS/PDF and LES/FDF model, flamelet generated manifolds (FGM) model, and the Irregularly Portioned Lagrangian Monte Carlo Finite Difference (IPLMCFD) methodology for scalable simulation of high-speed reacting compressible flows.
    No preview · Conference Paper · Jul 2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: The scalar filtered mass density function (SFMDF) methodology is implemented into the computer code US3D. The SFMDF is a subgrid scale closure and is simulated via a Lagrangian Monte Carlo solver. The US3D is an Eulerian finite volume code and has proven very effective for compressible flow simulations. The resulting SFMDF-US3D code is employed for large eddy simulation of compressible turbulent flows on unstructured meshes. Simulations are conducted of subsonic and supersonic flows. The consistency and the accuracy of the simulated results are assessed along with appraisal of the overall performance of the methodology.
    No preview · Conference Paper · Jul 2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: Simulation and modeling of turbulent flow, and of turbulent reacting flow in particular, involves solving for and analyzing time-dependent and spatially dense tensor quantities, such as turbulent stress tensors. The interactive visual exploration of these tensor quantities can effectively steer the computational modeling of combustion systems. In this chapter, we discuss the challenges in dense symmetric-tensor visualization applied to turbulent combustion calculation, and analyze the feasibility of using several established tensor visualization techniques in the context of exploring space-time relationships in computationally-simulated combustion tensor data. To tackle the pervasive problems of occlusion and clutter, we propose a solution combining techniques from information and scientific visualization. Specifically, the proposed solution combines a detailed 3D inspection view based on volume rendering with glyph-based representations—used as 2D probes—while leveraging interactive filtering and flow salience cues to clarify the structure of the tensor datasets. Side-by-side views of multiple timesteps facilitate the analysis of time-space relationships. The resulting prototype enables an analysis style based on the overview first, zoom and filter, then details on demand paradigm originally proposed in information visualization. The result is a visual analysis tool to be utilized in debugging, benchmarking, and verification of models and solutions in turbulent combustion. We demonstrate this analysis tool on three example configurations and report feedback from combustion researchers.
    No preview · Chapter · Jan 2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Acute care facilities are connected via patient sharing, forming a network. However, patient sharing extends beyond this immediate network to include sharing with long-term care facilities. The extent of long-term care facility patient sharing on the acute care facility network is unknown. The objective of this study was to characterize and determine the extent and pattern of patient transfers to, from, and between long-term care facilities on the network of acute care facilities in a large metropolitan county. We applied social network constructs principles, measures, and frameworks to all 2007 annual adult and pediatric patient transfers among the healthcare facilities in Orange County, California, using data from surveys and several datasets. We evaluated general network and centrality measures as well as individual ego measures and further constructed sociograms. Our results show that over the course of a year, 66 of 72 long-term care facilities directly sent and 67 directly received patients from other long-term care facilities. Long-term care facilities added 1,524 ties between the acute care facilities when ties represented at least one patient transfer. Geodesic distance did not closely correlate with the geographic distance among facilities. This study demonstrates the extent to which long-term care facilities are connected to the acute care facility patient sharing network. Many long-term care facilities were connected by patient transfers and further added many connections to the acute care facility network. This suggests that policy-makers and health officials should account for patient sharing with and among long-term care facilities as well as those among acute care facilities when evaluating policies and interventions.
    Full-text · Article · Dec 2011 · PLoS ONE
  • Patrick Pisciuneri · S. Levent Yilmaz · Peter Strakey · Peyman Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: The ``Irregularly Portioned Lagrangian Monte Carlo'' (IPLMC) [1] in LES/FDF is extended to include the Eulerian flow solver in a coupled manner. The resulting methodology is for LES of reacting flows on massively parallel platforms, and is intended for LES of turbulent reacting flows described by complex kinetics. The new solver provides much improved scalability over its predecessor for utilization of a higher number of processors. Sample results are presented of LES of non-premixed flames, along with scalability benchmarks. [4pt] [1] Yilmaz, S. L., Nik, M. B., Sheikhi, M. R. H., Strakey, P. A., and Givi, P., An Irregularly Portioned Lagrangian Monte Carlo Method for Turbulent Flow Simulation, J. Sci. Comput., 47(1):109--125 (2011).
    No preview · Article · Nov 2011
  • S.L. Yilmaz · N. Ansari · P.H. Pisciuneri · M.B. Nik · C. Ottis · P. Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: An overview is presented of recent developments in filtered density function (FDF) modeling and simulation in turbulent combustion including high speed propulsion systems. These are primarily the development of a new Lagrangian Monte Carlo solver for the FDF, and the implementation of this solver on Eulerian domains portrayed by unstructured grids. With these developments, it will be much easier to conduct large eddy simulation (LES) of turbulent reactive flows for complex geometrical configurations. © 2011 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
    No preview · Conference Paper · Jul 2011
  • S. Levent Yilmaz · Patrick H. Pisciuneri · Mehdi B. Nik · Peyman Givi
    [Show abstract] [Hide abstract]
    ABSTRACT: "Turbulence is the most important unsolved problem of classical physics". That was Richard Feynman decades ago, referring to a century old problem. Today, the situation is no different. Turbulent combustion, which deals with a fluid mixture reacting and mixing under turbulent conditions (as found in rockets, jet engines, power generators, car engines, furnaces,...), is harder still. While a solution that would satisfy a physicist is yet to be found, engineers all over the world are tackling the problem with computational modeling and simulation. There are a plethora of models for turbulence and combustion with a whole wide range of competing characteristics of applicability, accuracy, reliability and computational cost. Nowadays, reliability is the key feature required of such modeling (but, most often than not, sacrificed or oversight) for the design of environment friendly and efficient machines. There exists an unproven (but undeniable) direct correlation between reliability and computational cost. However, the era of sacrificing the former because one cannot overcome and afford the latter for a full scale engineering application is over, thanks to TeraGrid and other resources for open research coupled with relentless efforts of countless developers to provide software that runs faster and better. This project is one sampling of how these resources are utilized to overcome an important research problem. We take on the Filtered Density Function (FDF) for large eddy simulation (LES) of turbulent reacting flow, which is a novel and robust methodology that can provide very accurate predictions for a wide range flow conditions. FDF involves an expensive particle/mesh algorithm where stiff chemical reaction computations cause quite interesting, problem specific, and in most cases extremely imbalanced (a couple of orders of magnitude) computational loads. The authors present the Irregularly Portioned Lagrangian Monte Carlo[?], an advanced implementation of FDF based on a simple and smart parallelization strategy that is implemented via optimized solvers and high-level public domain parallelization libraries (eg. Zoltan[?]). The methodology and a discussion of the implementation is presented along with results and benchmarks on the TeraGrid (NICS/Kraken and PSC/Bigben). Scaling and speed up comparisons demonstrate that a conventional parallelization is unable to scale beyond a 100 processors, whereas the new implementation can efficiently utilize 1000 processors for the same size problem, and has enabled the FDF methodology to tackle ever larger problems.
    No preview · Article · Jul 2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Since hospitals in a region often share patients, an outbreak of methicillin-resistant Staphylococcus aureus (MRSA) infection in one hospital could affect other hospitals. Using extensive data collected from Orange County (OC), California, we developed a detailed agent-based model to represent patient movement among all OC hospitals. Experiments simulated MRSA outbreaks in various wards, institutions, and regions. Sensitivity analysis varied lengths of stay, intraward transmission coefficients (β), MRSA loss rate, probability of patient transfer or readmission, and time to readmission. Each simulated outbreak eventually affected all of the hospitals in the network, with effects depending on the outbreak size and location. Increasing MRSA prevalence at a single hospital (from 5% to 15%) resulted in a 2.9% average increase in relative prevalence at all other hospitals (ranging from no effect to 46.4%). Single-hospital intensive care unit outbreaks (modeled increase from 5% to 15%) caused a 1.4% average relative increase in all other OC hospitals (ranging from no effect to 12.7%). MRSA outbreaks may rarely be confined to a single hospital but instead may affect all of the hospitals in a region. This suggests that prevention and control strategies and policies should account for the interconnectedness of health care facilities.
    Full-text · Article · Jun 2011 · Infection Control and Hospital Epidemiology
  • [Show abstract] [Hide abstract]
    ABSTRACT: A discrete variable representation (DVR) implementation of an one-electron polarization model (OPEM) for characterizing (H(2)O)(n)(-) clusters is described. For the (H(2)O)(90)(-) cluster, evaluation of the energy and gradient using a suitable DVR basis sets is about a 2 orders of magnitude faster than corresponding calculations using a Gaussian orbital basis set The DVR version of the code has been parallelized using OpenMP to enable molecular dynamics (MD) simulations of large (H(2)O)(n)(-) clusters
    No preview · Article · Aug 2010 · Journal of Chemical Theory and Computation
  • [Show abstract] [Hide abstract]
    ABSTRACT: Reynolds-Averaged Navier-Stokes (RANS) simulations of Lean Premixed Combustion (LPC) of methane–air in a bluff-body stabilized combustor were performed with several widely used turbulent combustion methodologies in order to assess their prediction capabilities. The methods employed are the Eddy Dissipation Concept (EDC), the Composition Probability Density Function (CPDF) and the Joint Velocity–Frequency-Composition PDF (VFCPDF) models. Where needed, two different models were employed for turbulent transport closure, namely the Renormalization Group (RNG) k-ϵ and Reynolds Stress Transport (RSM) models. The combustion chemistry was represented by two separate augmented reduced mechanisms (ARM9 and ARM19) in order to assess the influence of chemical mechanisms on calculations. Mean temperature and major species predictions of all of the employed methodologies compared well with the experimental data. Intermediate and emission species predictions were sensitive to the resolution of turbulence viscosity, which changes the effective diffusivity of the species. NO emissions predictions were in error by an average ±5 ppm with the EDC models and the CPDF model, with the VFCPDF model showing a somewhat better prediction of NOx. Calculations for some intermediate species (especially H2) deviated qualitatively from the experimental data, which highlights some of the limitations of these methodologies commonly used in detailed prediction of emissions for various fuel blends.
    No preview · Article · Jul 2010 · Combustion Science and Technology
  • Source
    M. B. Nik · S. L. Yilmaz · P. Givi · M. R. H. Sheikhi · S. B. Pope
    [Show abstract] [Hide abstract]
    ABSTRACT: The joint velocity scalar filtered mass density function methodology is employed for large eddy simulation of Sandia National Laboratories' flame D. This is a turbulent piloted nonpremixed methane jet flame. In velocity scalar filtered mass density function, the effects of the subgrid-scale chemical reaction and convection appear in closed forms. The modeled transport equation for the velocity scalar filtered mass density function-is solved by a hybrid finite difference/Monte Carlo scheme. For this flame, which exhibits little local extinction, a flamelet model is employed to relate the instantaneous composition to mixture fraction. The simulated results are assessed via comparison with laboratory data and show favorable agreements.
    Full-text · Article · Jul 2010 · AIAA Journal