Upload
adam-torby
View
136
Download
5
Tags:
Embed Size (px)
Citation preview
Numeca advanced Developments for Better Products
A new wave in fluid dynamics
3
Numeca International Why Numeca?
Major Benefits to Your Organization• Improved performance, leading to a more effective
design process.
• Significant time and cost reduction throughout the whole CFD chain. Numeca grid generation automated systems automesh-4GTm, combined with the significant speed-up in cPu with its cPu Booster, guarantee a reduction of one to three orders of magnitude in your engineering and cPu time, on large test cases and com-plex geometries.
• Improve overall engineering efficiency and productivity.
• Improve decision support with reliable simulation.
• Effective design through optimization.
ServicesCustomer satisfaction is our main objective and we continuously improve our software and services to help you achieve successful product design. We offer a wide range of services including:
• Consultancy in a wide range of fluid, heat transfer, and multiphysics applications;
• Upgrades, advanced hot line and dedicated training and webinars;
• Flexible licensing arrangements to deal with all spe-cific organizational and business requirements;
• Extended partnership including:- Special arrangements to meet specific requirements
in terms of confidentiality, proprietary developments and know-how;
- Validation and calibration of our software to your specific test cases;
- Customization of the whole CFD chain towards customer specific requests;
- Integration of NUMECA software into customer desgin cycle chain;
- Access to software routines and R&D program through priviledged partnerships.
About UsNumeca has been providing computational Fluid Dynamics (cFD) Software, grid generation systems and consulting services worldwide since 1993. Numeca’s software systems are used for the simulation, design, and optimization of fluid flow and heat transfer. They are used by product developers and design and research engineers, allowing them to reach superior product qual-ity and performances, at a reduced engineering cost.
NUMECA International corporate headquarter is located in Brussels, with offices, resellers and service centers in USA, Germany, France, Russia, Italy, Spain, Poland, Japan, China, India, Indonesia, Malaysia, South Korea and Taiwan.
Application-Driven Grid Generation and CFD SoftwareNUMECA’s product strategy is based on the develop-ment of automated, integrated and customized soft-ware systems allowing optimal and rapid simulation, design and optimization. Our software closely follows industry requirements and needs:
• In grid generation, with AutoMesh-4GTM, covering the whole range of applications with tuned meshing solu-tions, as the pre-processor of most commercial CFD tools.
• In CFD, with FINETM/Turbo, FINETM/Open, and FINETM/ Marine, dedicated respectively to Turbomachinery, Aeronautics, Automotive, Multi-Physics and Marine applications.
• In design and optimization with FINETM/Design3D.
Advanced Development for Better ProductsNUMECA’s R&D team is a worldwide center of excellence comprising highly-skilled engineers and PhDs, of more than 20 nationalities, in Computer Science, Mathematics, Physics and Fluid Dynamics. NUMECA International par-ticipates in a large number of research projects with university departments, research laboratories and lead-ing industrial partners, allowing us to offer the latest breakthroughs in technology to our customers.
By choosing NUMECA you will gain access to the most advanced technology in the field of application-driven, fast and accurate CFD simulation software, automated full hexahedral mesh generation, solution-adaptive grid optimization, dedicated post-processing, CAD mod-eling, and optimization
2 3
contents
Products
• automesh-4G™ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
• FINe™/Turbo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
• FINe™/Open and FSI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9
• OpenLabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-11
• autoBlade™ & FINe™/Design3D . . . . . . . . . . . . . . . 12-13
• FINe™/marine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-15
• VNoise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-17
Key Features
• efficient aero-acoustics Prediction . . . . . . . . . . . 18
• Non-Deterministic Simulations . . . . . . . . . . . . . . . . . . . 19
• FINe™/Open for combustion
& Radiation modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-21
• multifluid & multiphase Flows . . . . . . . . . . . . . . . . . . . 22-23
• unsteady Phenomena Predicted
in Hours, Superior Solutions with
Non-Linear-Harmonic (NLH) approach . . . . . 24-25
• cooling Flow and Thermal effects
in Turbomachinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26-27
Numeca Solutions
• aerospace applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28-29
• automotive Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30-31
• Wind Turbine Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32-33
• Hydro engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34-35
© Julien DEBOCK 2010 © Julien DEBOCK 2010
© JUlIEN DEBOCK 2010
© JUlIEN DEBOCK 2010
© JUlIEN DEBOCK 2010
© JUlIEN DEBOCK 2010
4 5
automesh-4G™ New Generation of Innovative High-Quality Full-Hex and Hex Dominant Meshing Tools
for Whole CAE Applications
HEXPRESS™Unstructured—Full Hex- Non Conformal - body fitted grid generator for arbitrary geometries.No prism, no tetrahedra, no pyramid cells.
Customization to User-Specific Features:• Python based commands accessible through scripts• User-defined scripts for batch mode operations in design
process
HEXPRESS™/HybridIntegrated CAD cleaning and parallel grid generation system creating conformal body-fitted meshes on complex arbitrary geometry.
HEXPRESS™/Hybrid produces a 100-million cell grid on a full car including the underhood in 1 hour on 48 cores, starting from unrepaired STl files.
F1 racing car grid
HeXPReSS™ and HeXPReSS™/Hybrid use a similar volume-to–surface approach, suppressing the need for a surface mesh. Both systems run in batch or interactive mode and are interfaced to all major commercial cFD codes.
AutoGrid5™ combined with HEXPRESS™ enables the meshing of complex rotating
machine - Turbo Compressor
Complex hub/propeller interaction Courtesy ECN
Urban environment
AutoGrid5™Full automatic hexahedral block structured grid generator for all typesof rotating machinery and turbomachinery: complex axial, radial andmixed-flow configurations.
AutoGrid5™ generates grids in a few minutes, with just a few clicks through a wizard-based interface.
Axial multi-stage turbine
Counter-Rotating Open Rotor
Centrifugal multi-stage pump
Parametric design and meshing of cooled turbine blade.
Non axi-symmetric end walls
Automatic multistage mesh topology
Wind turbine mesh
IGG™Interactive Geometry Modeler and Multi-Block Structured Grid Generator for arbitrary geometry
• Simple and intuitive block creation tools• Automatic mapping and block connections• Automated meshing based on Phyton scripts • Meshing replay can be used in optimization loops.
Viscous layers around main plate and flap
Increase your Productivity by One Order of Magnitude
HeXPReSS™HeXPReSS™/
Hybrid autoGrid5™ IGG™
FINe™/Turbo
High Performance Computing
Advanced CFD Technology for Turbomachinery FINE™/Turbo HPC Capabilities Extend Beyond 1-Billion-Cell Grids on Thousands of Processors
FINE™/TurboFINE™/Turbo is an accurate and powerful block struc-tured Navier-Stokes CFD software dedicated to the simulation of internal, multi-stages rotating and tur-bomachinery flows.
FINe™/Turbo integrates:
• fully hexahedral automated grid generator AutoGrid5™ & IGG™;
• advanced density-based numerical algorithms with robust local preconditioning for incompressible flow;
• automatic performance curve construction & summary reports;
• easy-to-use and intuitive interface for fast solution set-up; • dedicated post-treatment for turbomachinery in CFView™; • batch process for transparent integration in a design cycle;• multi-physics: aero-acoustics, fluid structure interaction,
real gas modeling, particle tracking, cavitation, conju-gate heat transfer, etc.
Multi-stage axial compressor (Courtesy RR)
Automatic Performance Curve construction starting from design point within user-specified range
Integrated Nacelle & Rotating Fan
Counter Rotating Open Rotor
Propeller
Wind Turbine Centrifugal Compressor
Space Pump
Casing/hub treatment
Hydro Energy applications
Fan
76
cPu Booster:Convergence in 50 cycles
cPu BOOSTeR – axial compressor
• Allowing higher CFl number ~1000• Very fast convergence• Gain: 1 to 2 orders
of magnitude faster CPU time
FINeTm/Turbo HPc streamlines process:
• Fully transparent and virtual domain decomposition, allowing the launch of a grid with a small number of blocks on an arbitrary number of processors.
• Efficient and optimized automatic load balancing, taking advantage of a large HPC configuration (thou-sands of processors).
• Fully transparent reconstruction of the solution, allow-ing for the analysis and visualization of the solution on the original grid or at a coarser level in CFViewTM
Virtual Process Decomposition
Automatic Block Decomposition
Linear Speed-up up to 500 Processes
55 blocks split into 2640 virtual blocks (left) balanced on 1151 computational processes (right).
Stage 67 – Time reconstruction, snapshot of the solution at a given rotor position, close to tip, blade-to-blade
distribution of static temperature
Nasa Stage 67: 1 rotor, 1 stator, and recirculated injector, unsteady flow phemenomena simulated using the
Non-Linear-Harmonic approach, allowing the use of a 1-single-passage high-quality and high-density mesh.
Linear speed-up comparison: 3.5 Stage IDAC Steady Computation (80 million cells, 3 multigrid levels)
ApplicationStage 67 Transonic compressor
• 80 Million Cells with 3 multigrid levels• Non-linear Harmonic analysis with 4 frequencies• Computation completed on 512 processes in less than
1 day• Complex tip gap - recirculation channel flows
Preparation of the model in AutoGridTM using standard block topologies
Virtual decomposition following a “meta-block” tree structure
Reconstitution of the final solution onto the original block topology
Post-processing with CFViewTM
21 blocks
1000 blocks
9
FINe™/Open
Unstructured CFD Technologyfor Complex Flow Configurations
FINE™/Open FINE™/Open is an accurate unstructured Navier-Stokes solver, with extended turbulence and physical models.FINE™/Open solves any flow, from incompressible to low and high speed flows. FINE™/Open integrates HEXPRESS™ or HEXPRESS™/Hybrid.
Biomedical flows :Blood simulation in the aorta
Urban environment flows: Brussels, Courtesy Immobilière du Royal Rogiers and AM Art & Build/Montois Partners
Automotive applications
Convergence for complete car geometriesin 50 cycle! Jeep (19 million cells)
Solution-adaptive grid optimization for higher accuracy
Aerospace: transonic to hypersonic flows
FINe™/Turbo & FINe™/Open
are also available in one single
license.
MPPCI for Flutter Analysis
multiPhysics Fluid Structure Interaction
Modal Approach, Coupling through MPPCI and Strong Coupling in FINE™/FSI-Oofelie
Modal Approach for Forced Response Analysis
Transonic compressor blade magnitude of deformation on the blade surface [m] and relative mach number at mid span.
Fluid: water & Solid material: steel. Maximum enlargement of the slot:1.21 mm with no reinforcement, 0.069 mm with cylinders, 0.44 mm with plates.
FINE™/FSI-Oofelie for Curtain Header Displacement
Composite wind turbine blade displacement. FSI simulation: Coupling of FINETM/Open (CFD) and ABAQUS (Simulia-FEA) through MPCCI.
MPCCI: Mesh-based Parallel Code Coupling Interface from Fraunhofer Institute SCAI
MPCCI interface for codes coupling.
Rotor 67 Transonic compressor blade
Flutter on Agard Wing 445.6
Vortex Viwration Beam: Amplitude and Frequency Experiment and CFD Comparison
8
How? • Define and plug models from a GUI without the need
to worry about programming details or code structure
• A dynamic library is automatically built and loaded at
run time
• Compared to source-coded models, CFD solutions are
obtained with identical computing and memory costs
What are the benefits? Flexibility
• User-interaction with FINE™/Open
• Adaptation to your modeling requirements
cost/Time efficiency
• Fast and easy integration of your physical models
• Benefit from NUMECA’s CFD industrial environment
and features (HPC, parallelization, meshing capabili-
ties, advanced numerical methods)
• No programming knowledge is needed
• Free access for the entire FINE™/Open community
Example: Soot Formation modeling (Khan & Greeves)
OpenLabs
Customize and Add Your Own CFD and Physical Models
What Can Be Done?Free customization on
• Transport equations:
convection - diffusion - source equations
• Algebraic relations
• Source terms
• Diffusion coefficients
• Initial and boundary conditions
• Thermo-physical properties
(equation of state and transport properties)
• Output/post-processing
=>EQUATIONS@ PDE: Soot_Equation ->EXPRESSION: DDT(Ysoot) + CONV(Ysoot) = DIFF(YsootDiffCoeff) + SOURCE(SourceSoot) ->INITIAl_VAlUE: 0.0 ->TypeSolidBoundaryCondition: Neumann ->TypeInletBoundaryCondition: Dirichlet ->TypeOutletBoundaryCondition: Neumann=>AlGEBRAIC_DIFFUSION_COEFFICIENTS@ AlGDIFFCOEFF: YsootDiffCoeff ->EXPRESSION: EddyViscosity/PrSoot => SOURCETERMS@ SOURCE: SourceSoot ->EXPRESSION: B*Pressure*Xfuel*exp(-TA/Temperature) - A*Density*Ysoot*Turbk/Epsilon=>CONSTANTS@ CONSTANT: PrSoot ->Value: 0.7@ CONSTANT: B ->Value: 10@ CONSTANT: TA ->Value: 20230@ CONSTANT: A ->Value: 4.0
Flowchart of the process used to generate the OpenLabs library. The process is automatized and fully integrated
in the FINE™/Open environment.
OpenLabs GUI used to introduce the physical models. The library is built by a simple click in the GUI and automatically
loaded by the flow solver at run time.
OpenLabs meta-language sample for the specification of an unsteady boundary condition of the inlet absolute total pressure.
OpenLabs meta-language sample for soot formation modeling. A transport equation for the soot mass fraction with a nucleation
and oxidation source term is introduced with OpenLabs.
GUIResource-
fileText
Control-file
C++ code
Dynamiclib binary
Save the resource
-file
Compilation script
Control-fileCREATOR
These functionalities allow the addition and customiza-
tion of cFD and physical models for:
• Initialization of unsteady solutions; unsteady bound-
ary conditions
• Turbulence (e.g. turbulence modeling for a wind farm)
• Reactive flows; combustion; pollutant formation
• Radiation (e.g. modification of optical properties; take
soot into account)
• Multi-phase (e.g. cavitation)
• Heat-transfer; steady or unsteady heat sources
• Porous media
• And much more!
1110
Examples of Applications with OpenLabs
Example: Imposing an unsteady inlet boundary condition, GE-E3 bladeThe computation is set as 2D and non-rotational, and
the relative values are imposed at the inlet. To simulate
the effect of upstream weak of the vane, the formula
below is used to give the relative total pressure at blade
inlet, which is dependent on physical time t
=>AUXTERMS@ n=76@ omega=8283 @ P0-bg=223331.0 @ PI=4*atan(1.0) @ Period= 60.0/(omega*n) =>CUSTOM_BOUNDARY_CONDITIONS @ CUSTOMIZED_BOUNDARY_CONDITION: PtInlet ->EXPRESSION:P0-bg * (1.0 - 0.15 * pow((sin(n*tCoord/2+PI*Time/Period)),10))->ExistingBC: “Absolute Total Pressure” , row_2_flux_1_Main_Blade_upStream_inlet
Where:Po-bg relative total pressure at inlet, 223332.0 (Pa)n number of blades which is 76 Ω rotation speed, 8283rpmτ time period for one wake passage, τ=60 / (Ω*n)
A turbulent ethylene flame is simulated with the non-premixed combustion model and the P1 radiation model available in FINE™/Open. The soot forma-tion model by Khan & Greeves is introduced with OpenLabs. The soot mass fraction is shown in the figure.
GE-E3 blade with an unsteady inlet boundary. The time evolution of the static pressure over one period is observed.
12 13
autoBlade™ & FINe™/Design3D
3D Design and Optimization Rotating Machinery
Multipoint and Multi-Objective Optimization
AutoBlade™AutoBlade™ is an advanced and easy-to-use 3D para-metric modeler dedicated to the design of rotating and turbo-machinery blades including:• Conversion of CAD models to a fully parametric
definition,• Fitting module to “import” an existing geometry• Interactive graphical edition of the design parameters,• large variety of turbomachinery parametric models,• Parametric variables for:
• end walls;• non-axisymmetric hub/shroud; • blade profile; • splitter blades; • profile stacking; • technological effects;
• User defined parameters;• Dependency between parameters decreasing the size
of optimization parameters;• Tool analysis for blade and meridional contour;• Full undo/redo capability;• And much more!
Blade Re-cut Non-axisymmetric hub/shroud
User-defined parameters
AutoBlade Easy-to-Use
GUI including:
1. Model editor
2. Interactive
3. Contextual Menus
4. Customizable 3D view
5. View controls
6. Analysis tools
7. Undo / redo functions
Pre-defined parameter templates for various configurations: axial, centrifugal, radial, compressor, fan, turbine, pump,
wind-turbine.
FINE™/Design3DFINE™/Design3D is highly integrated 3D optimi-zation tool designed to improve the performance of rotating and turbomachinery blades. It allows designers to break the limit of traditional design rules and explore the concept of computer-based 3D innovative design.FINE™/Design3D integrates in a user-friendly interface, the 3D parametric blade modeler AutoBladeTM, genetic algorithms artificial intelligence, design of experiments techniques and efficient optimization algorithms.
Turbocharger innovative guide vane profile(“Courtesy BOSCH MAHLE Turbosystems”)
3D turbine multi-points optimization
Torque Converter optimization
KAPLAN Turbine optimization
3D Compressor blade multi-points optimization
Mixed flow fan stage optimizationNumber of blades and separation zones (red area)
are improved
Initial
Initial
Optimized
Optimized
ApplicationsFINE™/Design3D covers a large range of applications including multi-stage axial, radial and mixed-flow compressor, turbine, pumps, fans, wind-turbine or propellers.
14 15
FINe™/marine
Solutions for the Marine Industry
FINE™/MarineFINE™/Marine is a Flow Integrated Environment incor-porating High-Fidelity and Automated CFD Simulation dedicated to Naval Architecture. FINE™/Marine com-bines the powerful full-hex automated unstructured mesh generator HEXPRESS™, the incompressible RANS FINE™/Marine solver and the flow visualization and analysis system CFView™.
6 DOF Motion & grid deformation for accurate wave breaking capturing
15
Specific Interface Capturing Algorithms for Accurate Free Surface Resolution
Temporal mean of wave elevation – fixed DTMB 5415 in head waves (exp. by IIHR).
Mean of wave elevation – KCS ship (Tokyo CFD workshop 2005)
HEXPRESS™ for High Quality Full-Hex Complex Geometries Meshing
14
Success Story:
“For the last 8 years, we have been using 3 different CFD codes. FINE™/Marine is the first code that gave us confidence in the use of CFD tools, with results never differing much from experimental values. Meshing with HEXPRESS™, although not without problems in the case of really complex bodies (e.g. superstructures) is a pleasurable experience – much more advanced than common meshing tools.”
Dr. Piet Van Oossanen, Van Oossanen & Associates b.v.
Top:
details on mesh of stern appendages of
a hopper-dredger (Courtesy of IHC
Holland Dredgers b.v.).
Bottom:
details on mesh of aft ship of a generic
inland vessel.
Dedication – Accuracy – Integration for Naval Architects
Automatic Grid Adaptation
Sliding Grid for Hull - Propeller Interaction
Advanced Turbulence Modelling for Accurate Wake Capturing
Wind Effects
Efficient Real Scale Assessments
Some applications with FINe™/marine
Example of adaptive mesh refinement – Virtue Container Ship (Froude 0.272).
(Courtesy of Van Oossanen and Associates)
Oblique wedge impact.
Iso-contours of streamwise velocity-sliding grid computation on Hamburg test case with INSEAN propeller.
Model Scale (TOP) and Real Scale
(BOTTOM) CFD Result.
Real Scale on sea test result (LEFT) & CFD Result (RIGHT)
(Courtesy of IHC Holland Dredgers b.v.).
Isowake at propeller plane with Explicit Algebraic Stress Model (EASM) and grid refinement versus KRISO experiments.
Exp. CFD Comp.
Exp.
Comp.
Refined grid (3.88M cells)
Original grid (3.07M cells)
16 17
VNoise
Integrated CFD-Vibro/Aero-acoustic Solution
Integrated solutionA scalable tool for aero-vibro-acoustic analysis allowing simple pre-design calculations as well as detailed large scale analysis of coupled aero-vibro-acoustic problems over a network.
Key features:
• Vibro-acoustic • Aero-acoustic • Parallel/distributed solver• CAE Interfaces• Direct import of FINE/Open,
FINE/Turbo and NlH results
core Technologies:
• Direct and indirect BEM formulations, FEM solver, Multi domain and Mixed domain, Coupled FEM-BEM, Frequency response, Eigenvalues
• Full coupling with structural modal equations, Propagation with mean flow (BEM-FEM)
• locally and bulk reacting absorbers, Perforated tubes, Fast multi-domain analysis, Random vibroacoustics, Symmetries, load cases
• Post-processing: Tl (multi input multi output systems), Intensity, soundpower, Virtual impedance tube
Key functionalities:
• Easy integration in any CAE NVH chain• Advanced 3D interpolation for importing velocity
distribution and structural modes from the structural mesh into the acoustic mesh
• Fully automatic mesh coarsening coupled with advanced 3D interpolation for easy data exchange and manipulation
• Automatic shrink-wrapping for hole closure, fixing non conforming meshes, mesh size reduction, etc.
• Batch command files• Direct import of FINE™/Open and FINE™/Turbo results
Structural mesh of an automotive oil sump (left) as automatically processed by the Shrink-Wrapping algorithm in order to obtain a
suitable acoustic BEM mesh for sound radiation (right)
Intake system: Structural mode automatic projection from the structural FEM mesh (right) into the acoustic mesh (left)
Acoustic fatigue analysis for a Solar Array. Structural modes imported in VNoise (left) and used to start a coupled vibro-acoustic analysis with a diffuse acoustic source described with
VNoise principal component analysis feature. Results reported in terms of PSD and RMS acoustic quantities (right) as well as PSD of structural accelerations (middle) that are com-
pared with measurements in a semi anechoic chamber.
Sound Power radiated by a vibrating engine, computed
starting from measured vibra-tions. Results compared with acoustic measurements and another BEM code (Courtesy of the University of Kentuky)
Transmission loss analysis for an automotive muffler composed by three chambers, with perforated tubes (left).
Comparison between measured TL curve (blue) and VNoise results (red) is showed on the right.
Evaluation of the transfer function between a structural force applied on an oil sump and the radiated sound power. On the right is shown the acoustic mesh, and on the left the transfer function obtained by VNoise (blue) is compared with other numerical results (red). On the top is shown the case of a force applied in the vertical direction, while below the case
of a force applied horizontally
Random VibroacousticsPSD evaluation of accelerations and stress due to a diffuse acoustic source
Transmission Loss evaluation of complex mufflersBulk and locally reacting absorbing materials, perforated tubes, specific tools for Tl evaluation
Engine RadiationSound power evaluation and Velocity contribution Vectors for fast engine run-up analyses
Acoustic Transfer functionsStructural – acoustic transfer function evaluation for an automotive oil sump
Modular Approach for Aero-acoustics analysis:
Flow – Chart of a general aeroacoustic analys with VNoise. The modular approach with identification of source and propaga-
tion regions and the use of specific modules for each step makes complex analyses affordable in an industrial context
Tonal Noise radiation from a centrifugal FAN. CFD analysis was conducted on the rotating FAN and then the results
were imported in VNoise where the FWH modules was used coupled with the BEM solver
Import CFD results
Postprocessing
Define source and propagation area
Execute FEM/BEM analysis for
propagation region
Create acoustic mesh for source and propagation
region
Calculate tonal and/or non-tonal sources using FWH and/or
FlowNoise Courtesy of ElASIS
Re=3.106, Angle of Attack: 3 degrees Geometrical uncertainties:
• Uncertain relative thickness,
• Uncertain camberline.
Output: The airfoil pres-sure coefficient & its standard deviation.
NACA0012 Airfoil
Non-Deterministic Simulations
18 19
Managing Uncertainties and Risks within NUMECA’s CFD
Simulation Process
Operational and geometrical uncertainties in the CFD simulation:
• Tolerances of manufacturing
• Uncertainties on inlet and boundary conditions
• Model uncertainties
• ‘Incompressible’ numerical errors
• etc.
NASA Rotor 37
Operational uncertainty: Uncertain Inlet Total Pressure profile.
Pressure distribution around mid-span blade profile. Node locations (left).Shock locations are observed with larger uncertainty(right).Error bars represent variation of standard deviation σ.
Pressure ratio versus mass flow compressor map. larger uncertainty at low mass-flow.Error bars represent variation of standard deviation σ.
Operational uncertainty: Uncertain Outlet Static Pressure.Pitch-wise averaged pres-sure ratio radial distribu-tion. larger uncertainty near tip. (right)
Courtesy TU DElFT
Nearfield Acoustic Waves of Fan, Nacelle & Ground(NlH – 24 hours for 3 millions cells on 16 cores machine)
Efficient aero-acoustics Prediction
Ffowcs-Williams-Hawking (FWH) Approach Far field Noise Propagation
Counter-rotating open rotors
Permeable FW-H surface
Noise spectrum at observer location
Cost Efficient NLH for Noise Source & Nearfield Acoustics Generation
DucTeD FaN - Nacelle
OPeN ROTOR4 million grid points mesh11 hours on 6 processors (NLH computation)Similar unsteady simulation, based on sliding grids, takes 2 months on 200 processors with other methods
available in coming releases:• Ducting influence in propagation:
linearized Euler Equations (lEE) approach• Broadband noise via lES simulation
FINe™/Open for combustion & Radiation modelingRobust, Accurate and Reliable Modules for All Types of Combustion
Simulation of the Reactive Flow in a Combustion Chamber
Automatically Generated
full-Hex Mesh
RANS Flamelet
FGM Progress Variable
TU Darmstadt’s Generic Gas Turbine Combustor
> Non-Premixed combustion modeling: • Mixture fraction approach
• Tabulated chemistry method
• Enthalpy defect method for simulation of non-adi-abatic flames (radiative heat loss)
• Spray combustion
> Partially Premixed combustion modeling: • Mixture fraction/progress variable approach
• Two modeling methods available:
• Flamelet Generated Manifold technique
• Hybrid BMl/Flamelet model
> Premixed combustion modeling: • Progress variable approach for flame front tracking
• Zimonts‘ Turbulent Flame Speed Closure
Modeling the Combustion Process and the Radiative Heat Transfer in Furnaces
Full Hex HEXPRESS™ Mesh of an Aero
Engine Combustor
Easy to Use GUI for Combustion and Radiation Modelling
Non-premixed Combustion GUI
Partially-premixed Combustion GUI
Accurate and Reliable Predictions with the FGM Method
Sydney/Sandia Bluff-body
Stabilized Flame (Experiment
and simulated flow fields)
FGM - Flamelet Generated Manifold Approach: NUMECA ’s Unique Feature for Improved Modelling
2120
> applications
• Simulation of furnaces
• General non-premixed or partially premixed gaseous combustion processes
• Aero-engine combustors
• Gas turbine combustors
Attached Flame
Lifted Flame
> multi-species framework for general reactive flow simulations
• Finite rate chemistry
• Eddy dissipation modeling approach
> Radiation modeling:• First Order Spherical Harmonics Method (P1)
• Emission Model for optically thin media
• Finite Volume Method (FVM) for radiative heat transfer
• Weighted-Sum-of-Grey-Gases (WSGG) method for the determination of optical properties
> modeling Pollutant Formation:• NOx postprocessing module (thermal)
• Soot models: • One equation model of Greeves & Khan
• Two equation model of Moss & lindstedt
Combustion Look-up Table
FGM table created with TABGEN/Chemistry, NUMECA’s combustion table generation tool. The plot shows the
temperature manifold in dependence of the mixture fraction and the progress variable
z
r
z
r
z
r
FGM: Flamelet Generated
Manifold
Experiment Temperature OH Velocity magn.
This shows the streamlines coloured with the temperature of the simulated flow field in DLR Stuttgart’s model combustor.
The computation was carried out using Zimont’s modeling approach for premixed combustion processes.
Simulated temperature field in the IFRF glass melting furnace. The temperature field was obtained using a non-premixed
flamlelet method coupled with the P1 radiation model
Modeling of the Reactive Flow in Premixed Combustors
22 23
multifluid & multiphase Flows
TABGEN: User Generated Thermodynamic Table
for Multiphase & Multifluid Thermodynamic & Transport Properties
• TABGEN, Thermodynamic table generation tool for complex real Fluids and mixture definition based on the NIST-REFPROP database.
• 84 pure fluids & mixtures with up to 20 components: typical natural gas constituents, hydrocarbons, main air constituents, water, refrigerants, noble elements
User Defined Fluid Properties
Lagrangian Particle Tracking for Separators
Lagrangian particle tracking in turbo-machines
Multiphase flow – Lagrangian
particle tracking approach
one or two way coupling
Streamlines of relative particle velocities in a turbine (stator-rotor-stator configuration)
Perfect Gas, Ideal gas with Cp(T), Real fluid modeling
Modeling of Evaporating Sprays
Broken Dam Problem with the free surface VOF model. Evolution of the wave shape with time
VOF Model for Free Surface flow
Volumic Heat Sources for FIRE Simulation
Barotropic Model for Cavitating Flow
DELFT Hydrofoil Cavitating flow showing bubble detachment captured by FINE™/Turbo barotropic cavitation law. Courtesy TU Delft.
Cavitating flow on marine propeller. The wake is well captured by FINE™/Turbo barotropic cavitation law. Experimental results courtesy INSEAN Italy.
Compressible Cavitating Flows
Fire simulation with FINETM/Open volumic heat source model
Cavitating flow of R114 liquid on 4° Venturi (Thermo table approach). Compressibility effects are well taken into account here. Sharp discontinuity captured at the bubble frontier.Particle velocity streamlines
with fluid axial velocity con-tour field and iso-lines of par-ticle diameter at location of
complete evaporation.
Mean number diameter of initially
mono-dispersed evapo-rating droplet spray.
A Large Range of Models to Cover all CFD Applications and Physics
Porous Media for Flow in the Nuclear Reactor
25
unsteady Phenomena Predicted in Hours, Superior Solutions
with Non-Linear-Harmonic (NLH) approach
2 Orders of Magnitude Gain in CFD Turn-Around Time
NLH METHOD for Large Scale Multi-Stage Turbomachinery Unsteady Flow One passage mesh only: less memory and affordable CPU time
NLH Method for Casing Treatment(Courtesy lFA TU-München).
NLH for Mechanical Analysis
Direct mapping of pressure amplitude & phase for FEA analysis
24
NLH for Flow Distortion Flow distortion in external and internal flow
NLH METHOD for Clocking AnalysisOnly one run for the whole clocking spectrum
Stator 1 and Stator 2 are aligned
Solution after rotation of the stator 1 by 80% of its pitch.Efficiency is better at this clocking angle. Result (in black dot)is confirmed by full unsteady computation (in white dot).
Non Linear Harmonic (NLH) CPU Time – 3D Radial Turbine
2 orders of magnitude gain in CPU time. NlH compared to full unsteady with reference
to quasi-steady mixing plane in logarithmic scale.
Counter-Rotating Open Rotor near field acoustics signature
NLH for Aero-Acoustics
26 27
cooling Flow and Thermal effects in Turbomachinery
FINE™/Turbo, FINE™/Open & AutoMesh-4G™ Offer Flexibility in Handling Cooled Turbine & Conjugate Heat Transfer (CHT) FlowsFull meshing flexibility in automesh-4G™ by:• Parametric design and meshing in AutoGrid5TM;• Fully hexahedral unstructured meshing in HEXPRESSTM;• Extremely complex configuration meshing in HEXPRESSTM/
Hybrid.
Rich options for solving Turbine conjugate Heat Transfer by meshing real geometry or by source/sink terms without channel meshing.
Internal cooling passages, solid body and external blade main channel meshing (AutoGrid5™)
All Hexahedral Cooled Channel Structured Mesh
CHT
26 27 conjugate Heat Transfer simulation is made “simple” with non-matching interfaces between solid-fluid and fluid-fluid blocks for higher quality of mesh
Streamline and temperature in fluid and solid body. 1 stage (IGV + Rotor), 20 million cells on 45-core cluster.
CHT simulation compared to test result. Blade surface temperature. Mark II turbine test case
Local Source & Sink Terms for Film CoolingNo meshing of the channels is required
Fully Hexahedral Unstructured Mesh (HEXPRESS™)
cooling module Full mesh
Full mesh
cooling module
Increasing blowing ratio
experiment : red dotscooling/Bleed : blue continuous line
Full mesh : blue dotted line
FINE™/Open with HEXPRESS™ provide high quality full hexa unstructured resolution of cooled turbine phenomena.
Cooling Module and Full mesh flow configurations give similar
results. (Case Duden, 1999)
AutoGrid5™ allows easy positioning of local source & sinklocations: single hole, line of holes or slots
Cooled Turbine Parametric Design and Automatic Meshing in AutoGrid5™
Interactive and easy-to-use graphical user interface.
Extremely Complex Configuration Meshing in HEXPRESS™/Hybrid
NUMECA CFD Solutions for aerospace applications
Broad Range of Applications: External, Internal, Low Speed, High Speed,
Thermal, Fluid Structure Interaction
HEXPRESS™/Hybrid Hex Dominant Parallel Meshing of Hyper-Complex Full Aircraft Configuration in Hours
Extended Range of Capabilities for Aircraft Installation and Integration
Aircraft configuration study: Open Rotor-Fuselage interaction (top left), Pylon-Open rotor interaction
(top right), Ground effect at take-off (bottom left), Wing-body interaction (bottom right)
28
High-Lift Device Design
Alpha=31°
High-lift devices performance at low speed configuration
Fast and Accurate Prediction of Aerodynamics of Passenger Airliner
HVAC in an aircraft cabin
Passenger Comfort Analysis
Flutter Analysis
Flutter Analysis: FSI coupling of FINE™/Open and an FEA code through MPCCI
Lift vs.angle of attack
Alpha=10°
Automated Full-Hex Mesh for Quick CFD Turn Around & Accurate Refueling Simulation
HEXPRESS™ full-Hex mesh of a wing-box fuel tank for
refueling simulation
Hypersonic Flow
Very fast accurate aerodynamic performance of transonic, supersonic & hypersonic flows with CPU-Booster
29
CPU Booster: Convergence at 50 cycles.
31
NUMECA CFD Solutions for the automotive Industry
HydroplaningHydroplaning simulation of a tyre, employing free-sur-face and splash, plus particle tracking within the grooves
External Aerodynamics
Thermal Flow in a Cabin
Heat TransferFlow over a motor bloc
Breakthrough in CFD Solutionwith FINE™/Open• Full Second Order Accurate solution
• Agglomeration multigrid
• CPU Booster
Breakthrough in Full Automatic Meshing with HEXPRESS™/Hybrid
F1 (40 million cells)
NO CAD Cleaning - NO Surface Mesh
Hyper-Complex Configuration Meshing
30
Zooms on different parts of the buggy
HEXPRESS™/Hybrid allows meshing of complete car configurations ...
Buggy (24 million cells)
Standard approach:
Numeca’s solution:
Days / weeks of engineering time
Hours with HEXPRESSTM/Hybrid
Increase Your Productivity by at Least One Order of Magnitude
NUMECA CPU Booster combined with the agglomeration multigrid allows for running with CFL 1000 in around 50 cycles for complete car geometries!
This translates in a gain of at least one order of magnitude in convergence time!
Simulation of Terrain Effects on Wind Farm Energy Production with FINETM/Open
NUMECA Solutions
for the Wind Turbine Industry
Breakthrough in 3D Wind Turbine Meshing:Full automatic all hexahedra cellsmeshing in minutes with autoGrid5™
advanced applications:
• Complex configuration
• Cross wind in single blade passage
• Unsteady flow simulation
• Vibration
• Flow generated noise
Breakthrough in Flow SimulationComputing Time: Fully accurate solution in about 50 cycles with CPU Booster.
• Multi-point and multi-objective optimization
• Fully automatic process with no user intervention
• Optimization featuring Design of Experiments, Artificial Neural Network and Genetic Algorithms
• Powered by autoBlade™, 3D parametric blade modeler with pre-defined template for Wind Turbine
3D flow features at high wind speed
Black: Optimized BladeRed: Initial Blade
2.5 MW twist distribution optimized
blade with a gain in annual energy production
between 5 and 10%.
Fluid Structure Interactions• Strong fluid/structure interaction
with FINETM/FSI-Oofelie
• Modal approach in FINETM/Turbo
• Coupling to FEA commercial or
in-house tools through MPCCI
FSI - Enhancing the torsional flexibility of the blade by optimal selection of the composite fiber orientation
Wind speed prediction for wind turbine placement
in urban environment
32 33
Breakthrough in Full 3D Wind Turbine Blade Optimization
with FINE™/Design3D
Non Linear Harmonic Method: A Major Breakthrough in Unsteady CFD Simulations
• Accurate Rotor/Stator interactions
• Single passage mesh simulations
• Accurate unsteady solution with low number of harmonics (Blade Passing Frequencies)
• Reconstruction in time of the solution
Advanced CFD and Optimization Solutions for Hydro engineering
High Quality all Hexahedra Cells Meshing with AutoGrid5™ and HEXPRESS™
Structured mesh for Francis Turbine stay vane, guide vane and runner (AutoGrid5™ - mesh generated in a few minutes on a standard PC)
Unstructured mesh for spiral casing and distributor (HEXPRESS™ - < 1 hour CPU time for 1 million cells)
Before After
Full 3D Blade Optimization with FINE™/Design3D
Optimized blade and comparison of Blade Sections (initial in red & optimized in green)
Advanced Cavitation ModellingComparison of Non-linear harmonic and sliding grid simulation: instantaneous Blade pressure distribution and pressure fluctuation amplitude through FFT
3534
CPU cost comparison for Non-linear Harmonic and Sliding Grid simulation with reference to Steady Mixing Plane simulation.
Non-Linear Harmonic
Sliding Grid
Mesh size 900,000 15,000,000
RAM 1.5 Gb 7.5 Gb
Iterations 500 40,000
CPU Time w.r.t.Steady Mixing Plane
5 > 1000
Centrifugal Pump - Cavitation increase on suction side of the impeller with decreasing NPSH
NUMECA International
CORPORATE OFFICEChaussée de la Hulpe,189, Terhulpse Steenweg
1170 Brussels - BELGIUMTel: +32.2.647.83.11 - Fax: +32.2.647.93.98
email: [email protected] - http://www.numeca.com
NUMECA is distributed worldwide: please refer to our website for more details.