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: Sibendu Som Applicant's institution: ANL Applicant's division: ES Project Name: gm_gasexchange Project title: Large Eddy Simulation of IC Engine Flows Associated funding: LDRD funding for Argonne, GM funding for GM and Univ. of Michigan Other Systems: GM R&D Science: To enhance engine fuel efficiency and lower engine emissions, new strategies such as multiple injections, stratified charge direct injection, and high levels of exhaust gas recirculation offer great potential. However, these strategies also render combustion stability issues and cause higher cycle-to-cycle variations. High cycle-to-cycle variations can lead to misfire as well. This is undesirable from engine robustness, efficiency, and emissions standpoint. The cycle-to-cycle conditions are significantly impacted by engine in-cylinder flows generated during intake and exhaust processes. This is determined by geometries of the intake port, intake and exhaust valves and combustion chamber shape including piston bowl. Computational fluid dynamics (CFD) and large eddy simulation (LES) approaches are proposed to investigate the cycle-to-cycle variations. The specific objective is to determine the best practice for performing LES simulations of IC engine f lows, and to provide design guidelines for engineers to effectively use the LES tools to accelerate new engine designs. Project description: CFD with RANS (Reynolds-Averaged Navier-Stokes) approach has been applied to engine development for many decades and are still the main tool for engine design today. Due to solving ensemble or phase-averaged mean quantities, RANS is very much limited for average cycles, hence cannot capture cycle-to-cycle variations. The high fidelity DNS (direct numerical simulation) is too computationally intensive and is still not feasible for full engine simulations in the near future. By solving spatially filtered Navier-Stokes equation, LES has been highlighted to have the potential for predicting the details of the transient in-cylinder flows and offer the essential information regarding the nature of cycle-to-cycle variations a decade ago. However, the computational requirements and need for multiple-cycle simulations required by LES hindered its further research and application for IC engines. Together with great leap of computer power in past decade, Argonn e’s and GM R&D’s expertise in LES calculations, and availability of high-fidelity experimental data from University of Michigan (funded by GM R&D), this proposal aims at establishing the best engineering practices for LES calculations in IC engines. The increased predictive capability with LES will enable more efficient and robust engine design. GM R&D has developed many engineering CFD practices for RANS and LES simulations. One example is the numerical setup of valve opening and closing events for in-cylinder flow simulations. In the past, pending on the meshing topology used, a minimum valve lift together with a crank angle position is arbitrarily chosen to mimic the valve opening and closing events. In most cases, the minimum valve lift is determined mainly from solution stability and simulation cost considerations. This is true especially when the meshing topology involves non-orthogonal skewed meshes. A lack of accurate flow field measurements for validation in the valve vicinity also hinders progress in this area. Recent meshing advancements in CONVERGE in conjunction with the availability of accurate high speed PIV data render a new opportunity for such study. Using these toolsets, GM R&D has established a strategy for accurate predictions of engine in-cylinder flows for RANS simulations. The next ste p is to test this RANS methodology for LES simulations. With this proposal, the following main components of LES for engine flows will be employed: (1) Use dynamic structure sub-grid scale (SGS) turbulence model on unresolved small flow scales; (2) Use the best CFD setup derived from RANS simulation on minimum valve lift, opening /closing crank angle position and minimum mesh size in the valve seat region; (3) Use a 2nd-order convective scheme; (4) Evaluate model accuracy against available measured data from University of Michigan; (5) Explore different data analysis methods like POD on LES results. In summary, this proposal will be executed in a manner such that CFD simulations will collaborate closely with experimental diagnostics. The high-speed PIV data for the TCC-III engine from University of Michigan will be available for LES model validation. The ultimate goal will be to use LES for IC engine simulations for in-cylinder processes involving engine flow, spray, combustion, and soot emissions. The specific methods used for these flow simulations with moving boundaries will include: 1) Finite volume schemes 2) RANS turbulence models 3) CONVERGE software is being used - 2 base and unlimited child licenses available Industry partnership: General Motors - Research and Development Group Project URL: http://verifi.anl.gov/ Requested allocation: 500000 Q1: 200000 Q2: 200000 Q3: 50000 Q4: 50000 Justification: The simulations will be performed with CONVERGE code. As reported by all the other projects of the PI (Sibendu Som), the scaling efficiency is quite good (more than 70%) up to 2000 processors for typical engine problems Storage requirements: 1 TB 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