Fully Automatic In-cylinder Workflow Using HEEDS / es-ice / STAR-CD Simon Fischer – Senior ICE Application Support Engineer – CD-adapco Nuremberg Office Gerald Schmidt – Director/Powertrain - CD-adapco New York Office
Outline Introduction and Enabling Technology Workflow and HEEDS Portal Description – Underlying Scripts – Data Preparation – CAD Modification – Surface Preparation – Automatic Meshing – Physical / Numerical Model Setup – Running STAR-CD – Postprocessing Demo Study: The Optimized Race Engine Further Examples of Successful In-Cylinder and Related Studies
Introduction and Enabling Technology HEEDS is a well established MDX tool applicable to many CFD problems – Red Cedar Technology recently developed dedicated HEEDS portals for pro-STAR, es-ice, and STAR-CD making it easier to define workflows for in-cylinder simulation The key enablers of MDX for in-cylinder simulation are robust moving mesh technology and high fidelity physics capability – A major focus of es-ice development has been to provide a fully automated robust meshing technology with minimum user intervention – STAR-CD provides a big library of state-of the art in-cylinder relevant physics models (from multi-phase to combustion chemistry) This presentation shows the possibility of in-cylinder simulation using a fully automatic HEEDS/es-ice/STAR-CD workflow
Workflow Description Generic scripted process of STAR-CCM+ surface preparation, es-ice fully automatic meshing and model setup, pro-STAR model setup, STAR-CD cluster submit, and es-ice/pro-STAR postprocessing Minimum User Interaction to Prepare Input Data Process, parameter setting/variation, io-operations, postprocessing managed by HEEDS MDO employing the STAR- CCM+ portal and recent STAR-CD, es-ice, and pro-STAR portals Meshing and physics parameters and results output managed through HEEDS parameter tagging
Workflow Description: Underlying Scripts Generic scripted process of STAR-CCM+ surface preparation, es-ice fully automatic meshing and model setup, pro-STAR model setup, STAR-CD cluster submit, and es-ice/pro-STAR postprocessing
Workflow Description: Data Preparation Minimum User Interaction to Prepare Input Data – Surface preparation in STAR-CCM+ (GeometryPreparation.sim) (Geometry from CAD-Modeller, Imported CAD, Discretized Surface) – Naming convention – Orientation, Scaling, Valve and Piston position – CAD/surface repair (clean CAD, waterproof surface) Deliver Boundary Conditions in appropriately formatted files
Workflow Description: HEEDS Process Process, parameter setting/variation, io-operations, postprocessing managed by HEEDS MDO employing the STAR-CCM+ portal and recent STAR-CD, es-ice, and pro- STAR portals Meshing and physics parameters and results output managed through HEEDS parameter tagging
Workflow Details: CAD Modification Optional: Direct CAD Design Parameter (if exposed) access through HEEDS‘s dedicated Portals: here STAR-CCM+ Portal • Also available: • NX, CATIA, CREO,... Parametrized STAR-CCM+ Geometry Preparation sim-File
Workflow Details: Surface Preparation Input STAR-CCM+ Geometry Preparation sim-File • Automated Parts Based STAR-CCM+ • Surface Remesh using es-ice best practices (Optional) Automated Additional Static • Part Volume Meshing Export of es-ice -appropriate .dbs Surface file • and additional ccm volume meshes
Workflow Details: Automatic Meshing Underlying meshing script featuring latest es-ice automatic • meshing technology and best practice User input is limited to a minimum of mesh parameters, e.g. • mesh base size, prism layer thickness, etc. No expert knowledge of es-ice template mesh generation is • necessary Yet scripted meshing is open source and fully customizable for • expert users
Workflow Details: Automatic Meshing es-ice scripted meshing HEEDS portal to automatically tag parameters e.g. Make a mesh study Script allowing for Automatic Parameter Tagging to expose • parameters to a HEEDS design study (e.g. Make a mesh study)
Workflow Details: Automatic Meshing High quality moving mesh with NO user intervention
Workflow Details: es-ice Model Setup es-ice scripted model setup Boundary conditions • Initial conditions • Predefined best practice • setup scripts Cold flow • Mixture preparation • Combustion study • All scripts open source and • fully customizable All scripts allowing for • Automatic Parameter Tagging to expose parameters to a HEEDS design study
Workflow Details: pro-STAR Model Setup pro-STAR scripted model setup Liquid phase setup for • injector boundary conditions, liquid film, and spray physics Predefined best practice • setup scripts Cold flow • Mixture preparation • Combustion study • All scripts open source and • fully customizable All scripts allowing for • Automatic Parameter Tagging to expose parameters to a HEEDS design study
Workflow Details: Running STAR-CD Scripted cluster submit including all file handling (copying forth and back), monitoring jobs, etc., handling multiple computational resources
Workflow Details: Postprocessing es-ice postprocessing plugin HEEDS portal to automatically tag results Enabling postprocessing in HEEDS for an MDX study Automatic Parameter Tagging to expose Results to HEEDS
Workflow Details: Postprocessing Scripted call of pro-STAR engine animation script •
Workflow Details: Postprocessing Fully embedded postprocessing in HEEDS
Workflow Details: Project Management Fully integrated management of Processes (Cold Flow, • MixturePreparation, Combustion) and related Design Exploration Studies (e.g. • Single Baseline, Mesh Sensitivity, Geometry Optimization, ...) within the HEEDS Environment
Demo Study: An Optimized Race Engine The sample study features: Automatic CAD modification through STAR-CCM+ CAD modeler Automatic generation of moving mesh for intake and compression stroke for each design using es-ice Lagrangian spray, wall impingement, liquid film Tracking of fresh air, fuel, and residual through active species Turbulent flow field assessment with STAR-CD Evaluation of engine data Optimization of design through HEEDS Users touch neither the es-ice nor the pro-STAR GUI, leaving required expertise to a minimum
Demo Study: An Optimized Race Engine Design parameters: Piston bowl shape and injector direction Goal: Maximize fuel and maximize turbulence at the ignition point
Demo Study: Results
Demo Study: Sample Results All engine relevant postprocessing through HEEDS UI Baseline Best Design
Demo Study: Performance Model size: ~ 1 mio cells @ BDC (1 mm base size), time span 380 degrees crank angle (intake and compression stroke) Total turnaround time per design: ~ 3 hours – Preprocessing, meshing, and postprocessing on 1 CPU: ~ 25 min – Run on 48 CPUs: ~2.5 h Parallel design assessment through HEEDS – Here 4 designs in parallel ~ 45 min per design occupying 192 CPUs – 100 designs evaluated in 83 hrs License requirements: – 2 es-ice pre/post – 4 es-ice run – 164 HPC STAR-CD – 4 enablestar – 5 STAR-CCM+ – 1 HEEDS
Examples of Successful In-Cylinder and Related Studies Using HEEDS/es-ice/STAR-CD Transient port flow optimization (Pre-Sales demo – confidential) – Maximize tumble and trapped mass through port design optimization – Fully transient: 300 designs in 2 days on 64 cores Diesel piston bowl shape optimization (JLR – customer presentation) – Presented at last year‘s European ICE Workshop in Maranello Prechamber gas engine workflow (Support demo for Caterpillar) Gasoline spray calibration in a bomb experiment (Internal training) – Calibrate spray break-up model to match experimental properties Automatic testing of es-ice automatic meshing performance (Internal) – Extraction of best practices and sensitivity study on meshing parameters for optimum mesh quality Evaluation of DARS combustion libraries in STAR-CD box model (Internal) – Tool to explore combustion libraries for auto-ignition and laminar flame speed in CFD operating mode, valuable for library validation Turbulent Jet Ignition (TJI) combustion model validation and best practice study (Internal) – Combustion model assessment and calibration
The Application of HEEDS MDO to the Optimization of Diesel Piston Bowl Design, Paul Dunkley, Jaguar Land Rover (Presented in Maranello at the European ICE Workshop 2015) Paul D. of JLR stated that he can justify the investment in HEEDS purely because of cost savings due to process automation, even before running one single MDX study. Due to the advanc ancements ements in es-ice e autom omation ation transient ient full cycle le es-ice/S ice/STAR TAR- CD in-cylin inder der MDX is now possib ible le!
Piston Bowl Optimization Demo (Internal) NO SOOT
Prechamber gas engine workflow (Support demo for Caterpillar)
Recommend
More recommend
