September 2017 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Yearly Allocation Request for Denso_simulations
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Kaushik Saha Project Name: Denso_simulations Division: ES Project title: Modeling of flashing and non-flashing sprays in Eulerian and Lagrangian approaches Direct Injection Systems Associated funding: Department of Energy, Office of Energy Efficiency and Renewable Energy Other Systems: Science: This project will use a Eulerian-Lagrangian Computational Fluid Dynamics (CFD) software called CONVERGE to model direct-injection spray formation under the flashing and the non-flashing conditions to understand the effect of the change in the degree of superheat of the fuel when injected into the combustion chamber. Project description: The project will have two components. First part will deal with flash boiling phenomenon of internal and near-nozzle regions in the Eulerian framework. Both single and binary component flash boiling with ethanol blended with iso-octane and/or gasoline fuel. Internal nozzle flow simulations will be carried out with map file writing down the flow variables and species concentrations at the nozzle exit during the injection duration. The objective of the map file would be to provide the necessary boundary and initial conditions for the second phase of the simulations where the Lagrangian spray calculations will be carried out. Here also, both single and binary-component fuels will be tested. An in-house code has been developed to couple with CONVERGE in order to simulate the additional vaporization of the droplets due to flash boiling. Such capabilities are non-existent in most IC engine codes and hence, will be of utmost relevance to the industry. Industry partnership: Convergent Science (CSI) Project URL: http://verifi.anl.gov Current FY Hours Used: undetermined amount New FY Requested allocation: 500000 Q1: 175000 Q2: 150000 Q3: 100000 Q4: 75000 Justification: Our past publication (J. Kodavasal, K. Harms, P. Srivastava, S. Som, S. Quan, K.J. Richards, M. Garcia, “Development of stiffness-based chemistry load balancing scheme, and optimization of I/O and communication, to enable massively parallel high-fidelity internal combustion engine simulations,” Journal of Energy Resource Technology; JERT-16-1022, 2016) together with MCS and Convergent Science discusses the improvements to the Converge tool that has resulted in significant improvement in scaling. Currently, we are able to scale the code up to 4096 processors on Mira with about 70% scaling efficiency for a fixed mesh size. This was achieved due to the implementation of: (1) MPI I/O, (2) Improved Communication, (3) METIS load balancing scheme, (4) Development of a new chemistry load balancing scheme. Storage requirements: 5 TB. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for SEI
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Anh Ngo Project Name: SEI Division: MSD Project title: The effect of the transition metals on the solid electrolyte interphase in Li-ion batteries Associated funding: LDRD Other Systems: Science: Li-ion batteries (LIB) have become one of the most advanced rechargeable batteries as they shown great promise as power sources that can lead to the next electric vehicle revolution. Currently, rechargeable Li-ion batteries are key components of the mobile power sources for portable electronic devices. With high energy efficiency, Li-ion batteries may also be used in various electric applications including energy storage for renewable energy sources from wind and solar , however its large-scale applications are still hampered by narrow working temperature ranges and safety concerns under. For Li-ion batteries with cathode materials such as the layered LiMO2 where M is a transition metal (M = Mn, Co, Ni), M(II) ions are dissolved from the cathode and transferred by diffusion and migration toward the opposite anode through the thin separator. The experimental observation of the charge – discharge curves at the graphite electrode indicate that M(II) deposition causes a large reduction in the reversible capacity of the graphite anode which becomes worse with increasing M(II) concentration. The actual mechanism by which transition metal dissolution affects the solid electrolyte interphase (SEI) growth on the anode is still unclear and remains puzzling. In this study, we plan to address in detail several reactions of M(II) ions with major products of the SEI including lithium carbonate (Li2CO3), lithium fluoride (LiF), and lithium oxide (Li2O). The project will be carried out in as part of the CEES-II EFRC (Argonne) by theorists Anh Ngo (MSD), Hakim Iddir (MSD), and Larry Curtiss (MSD), who specialize in Density Functional Theory calculations. We will focus on following questions: 1. What is the impact of an M ion and its doping concentration on SEI structures? 2. What are oxidation/charge states of M ions in SEI structures? 3. Which mechanism causes capacity fading in Li ion batteries in the transition metal cathode materials? Such study will provide valuable information and we will be able to validate our theoretical results with the ongoing experimental efforts at Argonne. Project description: Description: Structure optimizations of Li2CO3, LiF and Li2O (~100 atoms) with different magnetic configurations will be carried out using (2 x 2x 2) bulk unit cells. By comparing different magnetic ordering energies we will conclude which configuration corresponding to the ground state energy. Once the ground state structure is found we will go on to address the SEI structural dependence of M ion doping concentration. We intend to carry out the density functional theory (DFT) calculations using the VASP package and the GPAW code. Periodic boundary conditions (PBC) will be imposed in order to maintain periodicity of the solid support. The gradient corrected PBE density functional will be used for these calculations. Those programs are currently available on Blues cluster and scale reasonably well on parallel processors. Since SEI doped M is a correlated complex oxide, we will perform density functional theory (DFT) calculations using the Heyd-Scuseria-Ernzerhof (HSE) hybrid functional and perform the atomic relaxations within the unit cells. Energy optimization of the SEI structure (~100 atoms) using HSE takes approximately 9216 CPU hours (16*6*96) and we plan roughly 55 such calculations. For doping calculations, each takes approximately 6912 CPU hours (16*6*72) and we plan to run 68 of these calculations. So our total request is 9216*55 + 16*6*72 *68 = 976896 core hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 970000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 220000 Justification: VASP performance on Blues is well established, we use the latest version maintained on Blues. Storage requirements: 1 TB is sufficient Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for Mullite
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Anh Ngo Project Name: Mullite Division: MSD Project title: The structure of disordered Mullite studied using a ab initio hybrid reversed Monte Carlo algorithm Associated funding: LDRD Other Systems: Science: Mullite is a material with an important role in the technology of classical and advanced ceramics for both traditional and advanced applications. The material is characterized by a very high thermal stability especially under oxidizing conditions, outstanding chemical stability in severe environments. Consequently, they are used in a great number of applications. The mullite structure can be written as Al2[Al2x+2Si2-2x]O10-x where x denotes the number of oxygen atoms missing per average unit cell and generally ranging between 0.2 and 0.5, x = 0.25 corresponds to 3:2 mullite, and x = 0.4 corresponds to 2:1 mullite. In mullite phases vacancies are introducing into the structures by replacing Si with Al while maintaining the overall charge neutrality . With higher alumina content, increasing x, the complexity of the structures increases with an increased number of O vacancies. The mullite structure becomes more and more disordered. Project description: Understanding the nature of the disordered structures and physical/chemical properties of mullite is therefore an area of increasing importance and has been the focus of investigation for many decades. In this project we would like to give a complete description of the real mullite. We will use ab initio energetic cluster expansion techniques to explore all distinct interactions/mechanisms between/of Al, Si, Oxygen and vacancy in the mullite. We then integrate these effective interaction parameters to build a fully microscopic model with our new developed ab initio hybrid reversed Monte Carlo. The project will be realized by collaboration between Anh Ngo (MSD, argonne) and Peter Zapol (MSD, argonne) and Ray Osborn (MSD, Argonne). Our results obtained from ab initio hybrid reversed Monte Carlo can be directly compared to diffuse scattering data measured at APS from Osborn’s group. To obtain our goal the following calculations will be performed: (a) The initial calculations will focus on finding the ab initio cluster expansion energies (CE) for all pair, triplet and higher order interactions. (b) To find real disordered structures, in our further investigations we will use CE obtained from step 1 to perform ab initio hybrid reversed Monte Carlo. This work will plan to use the high-performance computing facilities at LCRC. The first-principles calculations will be carried out using the electronic-structure code VASP. The Blues cluster is particularly suitable for this research. In order to obtain good Cluster expansion nergies we will plan to do 200 energy optimized calculations each calculation contains 60 atoms takes approximately 3072 CPU hours (16*4*48 ). The ab initio hybrid reversed Monte Carlo calculation takes roughly 26880 CPU hours (16*10*168) we plan to run 8 of these calculations. So our total request is 200*3072 + 8*26880 = 829440. In summary, the total requested allocations for this project is 829440 CPU hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 930000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 180000 Justification: VASP performance on Blues is well established, we use the latest version maintained on Blues. Storage requirements: 1 TB is sufficient Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for GNR-Au
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Anh Ngo Project Name: GNR-Au Division: MSD Project title: First-Principles Studies of the superlubricity on metallic substrate Associated funding: Strategic LDRD Other Systems: Science: Superlubricity is an emerging phenomenon where the friction force between the sliding materials is so small that it is almost vanished as compared to a usual stick-slip motion. Although friction is a macroscopic phenomenon, microscopic contacts are the key for the process. So far superlubricity has been demonstrated in macroscopic level including a recent success story of diamond scrolls on graphene by a CNM group [D. Berman et al, Science 348, 1118 (2015)]. Recently, Saw Hla’s group has performed a preliminary work for the lateral force measurement of individual Co-porphyrin molecules (TBrPP-Co) on graphene nanoribbons (GNR) grown on Au(111) surface. Comparison between the required lateral force of the same molecule on Au(111) surface and that on the GNR leads to an exciting finding: Almost zero lateral force is required to move the molecule on GNR while a relatively large force is needed to move the molecule on Au(111). This opens a novel research direction to i nvestigate molecular superlubricity on different low dimensional substrates such as silicene, borophene, phosphine and transition metal di-chalcogenides. The project will be realized by collaboration between Hla (CNM, argonne), an STM atom manipulation expert, Guisinger (CNM, argonne), a leader in synthesis of novel 2-dimensional materials, and the theorists Anh Ngo (MSD, argonne) and Peter Zapol (MSD, argonne), who specialize in DFT calculations of molecules on surfaces. Project description: To provide insight into electronic properties of molecular superlubricity we propose to perform a first-principles study for TBrPP-Co molecules adsorbed on graphene nanoribbons (GNR) grown on Au(111) surface. The following calculations will be performed: (a) The initial calculations will focus on finding the stable geometries of TBrPP-Co molecule adsorbed on graphene nanoribbons (GNR) grown on Au(111) surface. The total energies and electronic structure of several oriented structures of GNRs on Au(111) surface will be investigated . This will help improve our understanding of physical properties of molecules adsorbed on GNR grown on Au(111) surface . These theoretical results will be compared with the available STM experiment results. (b) In our further investigations we will calculate the total force as a function of the tip – Molecule distance. We will be able to validate our theoretical results with the ongoing experimental efforts at CNM, Argonne. This work will plan to use the high-performance computing facilities at LCRC. The first-principles calculations will be carried out using the electronic-structure code VASP. The Blues cluster is particularly suitable for this research because the studied systems are of large size, usually around 600-900 atoms. Structure optimizations of the system including a Co-porphyrin molecule deposited on GNRs on Au(111) surface (M/GNRs/Au (111)) with different magnetic configurations and GNRs widths will be carried out. By comparing different Co magnetic ordering we will conclude which configuration corresponding to the ground state energy. Once ground state structure is found we will go on to address the total force as a function of the different tip heights to extract lateral force information. Energy optimization of M/GNRs/Au (111)) (~800 atoms) using GGA+U takes approximately 43008 CPU hours (16*16*168 = 43008) and we plan roughly 17 of such calculations. For total force calculations. The calculation takes roughly 6915 CPU hours (16*6*72) we plan to run 37 of these calculations. So our total request is 16*43008 + 35*(16*6*72)= 930048 core hours. In summary, the total requested allocations for this project is 930048 CPU hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 950000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 200000 Justification: VASP performance on Blues is well established, we use the latest version maintained on Blues. Storage requirements: 1 TB is sufficient Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for IntensitySim
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Anh Ngo Project Name: IntensitySim Division: MSD Project title: Simulation Support for APS Data with ab initio Hybrid Reverse Monte Carlo Associated funding: LDRD Other Systems: Science: Our high-level goal is to improve the methods for determining large-scale structure for materials imaged at the APS by using Hybrid Reverse Monte Carlo, a combination of DFT, energy, and X-ray intensity calculations. Project description: Development history: The Calculate_I software came out of the Discovery Engines LDRD for combining experimental science with simulation. The program is written in Fortran with MPI for parallelism. OpenMP integration is planned. The code is compute-intensive: it reads a small amount of data produced by experiment and writes out a ~1 MB atomic structure dataset at the end. The computational method used models and improves a simulated atomic struction. A Monte Carlo iteration process then proposes improvements. The potential improvements are evaluated based on an energy calculation and an X-ray scattering intensity comparison to experimental data. The programming model used is Fortran + MPI. The code has been tested on LCRC scales and is expected to scale beyond that. We think that the analysis of data recently produced at CHESS by Ray Osborn will produce a Nature Materials paper. This project will use our software, Calculate_I, supported by 2000 VASP calculations. The plan to run VASP 2000 times and Calculate_I once for each experimental data set. The project initially contains the developers: Anh Ngo and Justin Wozniak. We will perform structure optimizations based on APS experiments of mullite and NaXV2O5 with different magnetic configurations will be carried out using (2 x 1 x 2) bulk unit cells. By comparing different magnetic ordering energies we will conclude which configuration corresponding to the ground state energy. We will determine the ground state structure as a result of the density functional theory (DFT) calculations using the VASP package. Periodic boundary condition (PBC) will be imposed in order to maintain periodicity of the solid support. Gradient corrected PBE density functional will be used for these calculations. From there, we can run Calculate_I to optimize the large scale structure. Those programs are currently available on Blues cluster and scale very well. Our total request is 1,000,000 core hours divided over the 4 quarters. Recent research products include: "Mechanical Flexure and Vibrational Maps of Individual Sexiphenyl Molecules on Ag(111)" Anh T. Ngo, Raymond Osborn, Justin Wozniak, and Peter Zapol Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 970000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 220000 Justification: We plan to use two codes: VASP and Calculate_I. The current version of Calculate_I scales well to 200 cores, which is the level at which we plan to run it. Storage requirements: 100GB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: InterfacialChem
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A new project on the LCRC cluster has been requested. Please forward the information on to the LCRC Allocation sub-committee. Applicant's name: Lei Cheng Applicant's institution: ANL Applicant's division: MSD Project Name: InterfacialChem Project title: Reactions at Electrolyte-electrode Interfaces Associated funding: JCESR Other Systems: Science: Events occurring at the electrolyte-electrode interfaces often have profound impact and sometimes even are limiting on the overall performance of the cells. For example, strong adsorption of certain species on an electrode surface might limit the surface accessibility for other active species, and in some cases cause undesired adsorption-induced chemical reactions (bond weakened by adsorption, similar to most catalysis mechanisms). The dispersion of nucleation centers governed by the strength of the adsorbate-surface interactions determines particle growth in precipitation—dissolution chemistry, which in turn controls the kinetics of the reverse discharge/charge reactions in chemical transformation batteries. In some intercalation battery systems, the desolvation/intercalation at the surface layer is the rate-limiting step and the surface rearrangement associated with the process will further impact the subsequent steps and cycles. accuracy. Project description: Density functional theory (DFT) provide useful tools to study the phenomenon mentioned above, especially when chemical reactions are involved. The inclusion of explicit solvent molecules in the model is critical as all components might play active role in altering chemical and physical pathways, thermodynamics and kinetics of these processes. However, sampling solvation structures close to solid surface using ab initio method is computationally costly. In this work, we propose using computationally inexpensive classical potentials (force-fields) to sample atomic configurations at the interface, and solvent configurations under electric field; these will be followed by structure optimizations using DFT to compute thermodynamics, and Nudged Elastic Band (NEB) method to calculate kinetics of the surface events with chemical accuracy. Systems to be studied include, but not limited to: (1) how the nature of electrolyte and cathode affect the desolvation/i ntercalation process at the electrolyte/cathode interface by systematically varying salt and solvent species, concentration (dilute vs highly concentrated), and cathode materials, (2) degradation of various electrolyte species adsorbed at cathode and anode surfaces before and after charge transfer, (3) the effect of electrode surface modification/functionalization on active species adsorption and charge transfer, (4) discharge product particle distribution/dispersion on the cathode surface of chemical transformation systems and the impact on the growth process. VASP will be used for most DFT calculations. Industry partnership: Project URL: Requested allocation: 400000 Q1: 100000 Q2: 100000 Q3: 100000 Q4: 100000 Justification: Storage requirements: The requester has used undetermined amount hours of their initial startup project. In addition to approving an initial amount, please specify a Category and Subcategory for this project. For a list of the current selection of approved categories, please see: https://wiki.lcrc.anl.gov/wiki/Processes/Categories Once the Allocation committee has approved the project, please go to the Project Management page to create it: https://accounts.lcrc.anl.gov/projects.php Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for CAT-engine-modeling
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Sibendu Som Project Name: CAT-engine-modeling Division: ES Project title: Development of sub-grid turbulent combustion models for LES under engine conditions Associated funding: DOE Office of vehicle technologies under the small business voucher program Other Systems: Science: This project will use the computational fluid dynamics (CFD) software CONVERGE to simulate cycle-to-cycle variation (CCV) in spark-ignited gasoline engines. The primary objective of the proposed research is to understand the impact of model settings, and simulation strategies on capturing CCV in spark-ignited (SI) gasoline engines through large-eddy simulation (LES)-based CFD. The ultimate objective is to develop best practices for industry-users of CONVERGE to run CCV simulations that accurately characterize the variation in the SI engines they intend to optimize, while also being computationally efficient. Project description: In the previous year, we worked with Convergent Science, Inc. (CSI), the makers of CONVERGE to identify the engine dataset to simulate and the version of CONVERGE to benchmark for this project. We are using a dataset from spark-ignition engine experiments on a Ford single-cylinder, direct-injected engine at Argonne. We performed multiple cycles of LES for this engine running on the LCRC cluster, Blues to simulate CCV, first using a base mesh size of 4 mm. We optimized various settings for the turbulence model, adaptive meshing and embedding, and solver settings, based on computational bottlenecks identified through preliminary simulation. We are currently running simulations with finer mesh resolutions (i.e. higher fidelity) of 2 mm and 1 mm base mesh sizes as well to understand the impact of mesh resolution on capturing CCV. In FY18, we intend to explore the impact of introducing numerical perturbations in the simulations through multiple restarts, to evaluate whether this artificial numerical perturbation adversely affects the overall CCV prediction from multiple consecutive cycles. This is of importance for industry users of CONVERGE who might have restrictions on maximum walltime a simulation can run, before it needs to be restarted. Further, we also intend to explore the potential of capturing CCV through multiple concurrent simulations, each with an initial perturbation introduced through a synthetic turbulent field. We would like to evaluate whether this concurrent simulation approach yields similar CCV values as running multiple cycles sequentially, as the concurrent approach has the potential of reducing overall turnaround time for CCV simulations by orders of magnitude on leadership-scale systems like Theta and Aurora. Industry partnership: Convergent Science, Inc. Project URL: http://verifi.anl.gov/ Current FY Hours Used: undetermined amount New FY Requested allocation: 800000 Q1: 200000 Q2: 200000 Q3: 200000 Q4: 200000 Justification: Our recent publication (J. Kodavasal, K. Harms, P. Srivastava, S. Som, S. Quan, K.J. Richards, M. Garcia, “Development of stiffness-based chemistry load balancing scheme, and optimization of I/O and communication, to enable massively parallel high-fidelity internal combustion engine simulations,” Journal of Energy Resource Technology; JERT-16-1022, 2016) together with MCS and Convergent Science discusses the improvements to the Converge tool that has resulted in significant improvement in scaling. Currently, we are able to scale the code up to 4096 processors on Mira with about 70% scaling efficiency for a fixed mesh size. This was achieved due to the implementation of: (1) MPI I/O, (2) Improved Communication, (3) METIS load balancing scheme, (4) Development of a new chemistry load balancing scheme. The above changes are not only expected to benefit calculations on Mira but also help on computing clusters like Bebop and Blues. Storage requirements: 5 TB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for cfd_enginemodeling
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Sibendu Som Project Name: cfd_enginemodeling Division: ES Project title: Numerical Simulations of Knock in spark-ignited engines Associated funding: DOE – Office of Vehicle Technologies under co-optima program Other Systems: Science: This project will use a Computational Fluid Dynamics (CFD) software called CONVERGE to model knocking combustion in spark-ignited engines. The primary objective of the proposed research is to develop turbulent combustion models which can accurately predict knock in a CFR engine. This objective will be achieved through a collaborative effort by combining computational diagnostics, combustion modeling, reduced kinetic mechanisms, Reynolds-Averaged Navier Stokes (RANS) simulations and LES. The model development will be primarily based on the G-equation framework for tracking the flame front and homogeneous reactor multi-zone model for capturing end-gas auto-ignition. Project description: In the previous year, a methodology which used G-equation and multi-zone combustion models was developed in the RANS context. The idea is to employ this hybrid approach to capture two distinct combustion phenomena: turbulent flame propagation and end-gas auto-ignition. The project for the current year is divided into two parts: In the first part, the developed combustion model will be employed to study fuel-engine interactions at engine-relevant highly boosted conditions. Numerical tests with a variety of gasoline fuel surrogates will be carried out to understand the implications of fuel chemistry and engine operating conditions. Parametric sweeps of compression ratio and boost pressure will be performed. In the second part of the study, the combustion model will be employed in the LES context to study the effect of cycle-to-cycle variations on knocking tendency. It is expected that the new methodology will lead to improved predictions of engine knock. Industry partnership: none Project URL: http://www.transportation.anl.gov/engines/multi_dim_model_home.html Current FY Hours Used: undetermined amount New FY Requested allocation: 2000000 Q1: 500000 Q2: 500000 Q3: 500000 Q4: 500000 Justification: Our recent publication (J. Kodavasal, K. Harms, P. Srivastava, S. Som, S. Quan, K.J. Richards, M. Garcia, “Development of stiffness-based chemistry load balancing scheme, and optimization of I/O and communication, to enable massively parallel high-fidelity internal combustion engine simulations,” Journal of Energy Resource Technology; JERT-16-1022, 2016) together with MCS and Convergent Science discusses the improvements to the Converge tool that has resulted in significant improvement in scaling. Currently, we are able to scale the code up to 4096 processors on Mira with about 70% scaling efficiency for a fixed mesh size. This was achieved due to the implementation of: (1) MPI I/O, (2) Improved Communication, (3) METIS load balancing scheme, (4) Development of a new chemistry load balancing scheme. The above changes are not only expected to benefit calculations on Mira but also help on computing clusters like Bebop and Blues. Storage requirements: 5 TB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for LES-sprays
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Sibendu Som Project Name: LES-sprays Division: ES Project title: Modeling dynamic coupling of internal nozzle flow and spray formation for Gasoline Direct Injection Applications Associated funding: DOE - Office of Vehicle Technologies Other Systems: Convergent Science Inc. Science: The effect of needle movement and internal nozzle flow on external spray formation have always intrigued the scientific community. The coupling of nozzle flow and spray formation is dynamic in nature and high-fidelity simulations with pragmatic choice of spatial and temporal resolutions are needed to elucidate the physics of this multiphase problem in Gasoline Direct Injection (GDI) systems. Such developments are of extreme interest to Department of Energy (DOE) and the automotive industry. Project description: The dynamic coupling of nozzle flow and spray formation will be achieved using Eulerian-Lagrangian Spray Atomization approach. The liquid fuel will remain in the Eulerian framework while exiting the nozzle. Thereafter, depending on local instantaneous liquid concentration in a given cell and amount of liquid in the neighboring cells, part of the liquid mass will be transferred to Lagrangian framework in the form of droplets. Such approach requires additional transport equation apart from the conservation equations of mass, momentum, species, energy, and turbulence in Eulerian framework. This additional equation is termed as the Σ equation. Σ represents the liquid-gas interfacial area per unit volume in a given computational cell. This special type of modeling in Eulerian framework is known as Σ-Y approach, which originated from flame surface modeling in the spray combustion community. This hybrid approach will have the potential of capturing the trans ient flow characteristics due to needle movement transcending downstream and affecting time-fluctuating spray phenomenon. The back-flow of chamber gas into the stepped holes of a GDI system causing the spray to taper towards the injector exit and our simulation work was one of the first few studies that demonstrated the phenomenon. Different chamber conditions will be used (flashing and non-flashing) will be used along with a variety of fuels (single and binary). The goal will be to provide an unforeseen development in GDI spray literature. Industry partnership: Convergent Science Inc. Project URL: http://www.transportation.anl.gov/engines/multi_dim_model_home.html Current FY Hours Used: undetermined amount New FY Requested allocation: 1000000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 250000 Justification: Our recent publication (J. Kodavasal, K. Harms, P. Srivastava, S. Som, S. Quan, K.J. Richards, M. Garcia, “Development of stiffness-based chemistry load balancing scheme, and optimization of I/O and communication, to enable massively parallel high-fidelity internal combustion engine simulations,” Journal of Energy Resource Technology; JERT-16-1022, 2016) together with MCS and Convergent Science discusses the improvements to the Converge tool that has resulted in significant improvement in scaling. Currently, we are able to scale the code up to 4096 processors on Mira with about 70% scaling efficiency for a fixed mesh size. This was achieved due to the implementation of: (1) MPI I/O, (2) Improved Communication, (3) METIS load balancing scheme, (4) Development of a new chemistry load balancing scheme. Storage requirements: 5 TB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for pozdyn
by accounts＠lcrc.anl.gov 05 Sep '17

05 Sep '17

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Mark Pfeifle Project Name: pozdyn Division: CSE Project title: Dissociation Dynamics of Primary Ozonides in Alkene Ozonolysis Associated funding: DAAD (German Academic Exchange Service) Other Systems: Linus (CSE) (no fixed time allocation, depending on current total load) Science: Ozonolysis is a major tropospheric degradation process for alkenes. In the first reaction step, a cyclic primary ozonide (POZ) is formed which is subsequently decomposed into a Criegee intermediate (CI) and a carbonyl compound. The CI population is initially hot due to the exothermic POZ formation. This nascent energy distribution of the CI is essential to adequately model the yields of promptly decomposing and stabilized CI. These quantities in turn are of considerable interest because the thermalized CI acts as an atmospheric oxidant for various species and enhances aerosol formation. It is commonly assumed that the excess energy is distributed statistically among the fragments in the spirit of phase-space theory. However, this assumption is questionable when the exit channel potential energy surface is strongly repulsive, as is the case with POZ decomposition. In this project, direct dynamics calculations for C2H4 + O3 are carried out. Besides concerted POZ cleav age, a step-wise diradical pathway contributes comparably to the reaction flux. Trajectory calculations, initiated at both exit transition states, serve as a basis to test the statistical partitioning assumption and to provide appropriate input distributions for CI master equation calculations. Project description: So far, direct dynamics calculations using density functional theory (DFT) have been successfully carried out for the concerted and step-wise primary ozonide decomposition pathways. Significant non-statistical effects have been found. However, the reliability of DFT especially for the step-wise channel is uncertain because of the multi-reference character of the diradical structure. In the next stage, we want to treat the step-wise pathway with strongly correlated electronic structure methods. Such methods are expected to be more accurate than DFT. Promising candidates are multi-configurational 2nd order perturbation theory (MCPT2) and restricted active-space 2nd order perturbation theory (RASPT2), which are implemented in Molpro and Columbus, and particularly multiconfiguration pair-density functional theory (MC-PDFT), implemented in openmolcas. Before running actual trajectories, the accuracy, performance and memory usage of the different methods and program packages will be thoroughly investigated. The accuracy can be tested by carrying out a series of energy calculations on the DFT minimum-energy pathway between the transition state and products and comparing the results with a reference calculation (e.g. multi-reference configuration interaction, MRCI). Parallelization is another topic of interest, even though the problem of Monte Carlo simulations itself is embarrasingly parallel, because several individual trajectories can be independently run on multiple cores on one node. The trajectory integration will be carried out with a locally modified version of DiNT (http://tcg.cse.anl.gov/papr/codes/dint.html) where the computational cost for this step is negligible compared to the quantum-chemical energy and gradient calculations within Molpro, Columbus, or openmolcas. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 200000 Q1: 200000 Q2: 0 Q3: 0 Q4: 0 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0