for CFD ge? February Arrival of cup of coffee Lunch break ROUND POSITION - PART 2 Professor Spencer London Frankfurt Coffee break afterwards.

Transcription

1 2nd Cluster Workshop Vehicle Physics Simulation Driven Design Black box or detailed knowledg ge? AGENDA for CFD 23 February :00 WARM-UP Arrival guests welcoming m over a cup cfee 9:30 INTRODUCTION ROUND Introduction to workshop 9:40 TAKE THE POLE POSITION - PART 1 KEYNOTE: Spectral/hp element, scale resolving modelling for high Reynolds number F1 Aerodynamics Pressor Spencer Sherwin - Imperial Collegee London Bringing Complexity CFD into CAD World an Engineer Boris Marovic - Mentor Graphics (Deutschl) GmbH Efficient methods for parametric shape optimization flow channelss Ioannis Nitsopoulos - ISKO engineers AG 10:55-11:30 PIT STOP Cfee break 11:30 TAKE THE POLE POSITION - PART 2 The influence mesh characteristics on OpenFOAM simulations DrivAer model Grigoris Fotiadis, Vangelis Skaperdas, Aristotelis Iordanidis - BETA CAE Systems S.A. 12:20-13:45 PITT STOP Lunch break 13:45 TAKE THE POLE POSITION - PART 3 On Need for Expertise when Using CFD - a Teacher s View Ulrich H. Rist - Institute Aerodynamics Gas Dynamics Universityy Stuttgart Below Glossy Surface: Methods, Errors Complexity in CFDD Pr. Gabriel Wittum - G-CSC, University Frankfurt 14:35-15:00 PITT STOP Cfee break 15:00 WIN THEE RACE Informationn about available subsidiess Presenting discussing submitted project appetizers 16:00 CHEQUERED FLAG End workshop ree will be ample opportunities for f furr technical discussions afterwards. Sponsored by From simulations to movement.

2 ABSTRACTS Pr. Spencer Sherwin Imperial College London Spectral/hp element,, scale resolving modelling for high Reynolds number F1 Aerodynamics The use computational tools in industrial flow simulations is well established. As engineering design continues to t evolve becomee ever more complex re is an increasing dem for more accurate transient flow simulations. It I can, using existing methods, be extremely costly in computational terms to achieve sufficient accuracy in se simulations. Accordingly, advanced engineering industries, such as F1 industry, is looking to academia to develop next generation techniques which may provide a mechanism for more accurate simulations without excessive increases in cost. Currently, most established methods for industrial flow simulations, including F1, are based upon Reynolds Averaged Navier- heart most commercial codes. There is Stokes (RANS) equations which are at naturally an implicit assumption in this approach a steady state solution. In practice, however, many industrial problems involve unsteady or transientt flows which RANS techniques are not well equipped to deal with. In order to refore address increasing dem for more physical models in engineering design, commercial codes do include unsteady extensions such as URANS (Unsteady RANS), Detached Eddy Simulation (DES). Unfortunately even on highh performance computing facilities se types computational models require significantly more execution time which, to date, has not been matched with a corresponding increase in accuracy a level sufficient to justify thiss costs. Particularly when considering computing restrictions F1 rules impose on race car design. Alternative transient simulation techniques have been developed d within research academic communities over past few decades. These methods have generally been appliedd to more academic transient t flow simulations with a significantly reduced level turbulence modelling. As industrial dem for transient simulations becomes greater computer "power per $" improves, alternative computational techniques, not yet widely adopted by industry, are likely to provide a more cost effective tool from perspective computational time for a high level accuracy. In this presentation we will outline dems imposed on computational aerodynamics within e highly competitive F1 race car design d discuss t next generation transient flow modelling that industry is looking to impact on this design cycle. 2

3 Boris Marovic, Mentorr Graphics (Deutschl) GmbH Frankfurt Bringing Complexity CFD into CAD World an Engineer Phrasess such as Simulation Driven Design Frontloading finding increasing use in numerical simulation environment. However actual design environment is far from being driven by simulation. In most cases simulation in such a scenario is slowing down design. With 3D CAD environment design changes can be created every few minutes no traditional CFD stware can provide that fast answerss as to drive next design any time close to design cycle times. With Meshing most CAD generated models (not idealized models) taking up at least several man-hours (manual meshing) but in most cases rar man- is abruptly interrupted if it were to wait for next design suggestions based on simulation results. Design product development engineers are creating new product designs, shapes structures ten want to days if not even weeks, design cycle know if ir design works or what where y should improve. They also don t have numerical knowledge CFD experts to underst what residuals, Y+ a broad range turbulence models have to be considered for good result accuracy. Often solution is praised as scripted or App-lified simplifications traditional CFD tools for those engineers. In I such a case experts define scripts or create a simplified GUI for engineer to enter ir boundary conditions get results automatically. Those however are ten scripts or Apps very limited in capability orr application don t fer any flexibility. This presentation will show a different approach o a CAD embedded CFD solution which utilizes a special meshing technology that enables automatic meshing most complex geometry solver technologies that reduce t user input to ir typical boundary conditions do not require any turbulence model selection or tweaking in order to reach r solver stability. Designs can refore be analyzed in hours compared to t days orr weeks. Bringing engineers much closer to ir desired design cycle compared to traditional tools but does not make m obsolete. 3

4 Ioannis Nitsopoulos ISKO engineers AG München Efficient methods for parametric shape optimization flow channels An important requirement in design optimization process is variation geometry in each loop. Until now this was only possible in a limited way in an automated process, since optimization needs a robustt flexible processs which uses potential existing design space. This presentationn provides an overview shape changing methods based on CAD network-based models introduces a new efficient practical method. In lecture entire procedure, from geometry description, on integration CFD simulation in an optimization process, through to evaluation results is shown by example a flow channel. 4

5 Grigoris Fotiadis, Vangelis Skaperdas, Aristotelis Iordanidis BETA CAE Systems S.A. Greece The influencee mesh characteristics on OpenFOAM simulations DrivAer model experimental results. Conclusions are derived with respect to importance mesh, optimum pre-processing strategy that ensures robust automation as well as high fidelity CFD simulations with OpenFOAM.. Computational Fluid Dynamics is a sophisticated, continuously exping field, providing its benefits to an ever increasing number engineers across industry. It fers great insight to complex flow problems that would orwise be difficult, if nott impossible, to gain. It has refore become an integral part complete design processs in all high-tech industries. In this study external aerodynamics CFD simulations are performed using OpenFOAM on three variants DrivAerr model, a realistic geometry with details representative current automotive designs. A thorough examination effect different meshess on solution convergence accuracy is performed. These meshess differ in terms f generation processs time involved, ir t density ir quality. Different meshing approaches are followed using pre- hybrid penta tetra meshes to hexa dominant polyhedral ones. Or factors considered are steady or processor ANSA, ranging from stard transient approaches, as well as importance including wind tunnel in simulationss to exactly match CFD 5

6 Bettina Lvogt scapos AG Schlosss Birlinghoven Sankt Augustin Using MpCCI to Solve Coupled CFD Problems arising in Automotive Design Numerical simulation has become a widely accepted established state- dynamics, especially in automotive design development process. However, re are some applications where fluid flow cannot be simulated realistically without taking or aspects into account: deformation spoilers orr wings, for example, influences fluid flow around car, which in turn influences deformation spoiler. For se so-called fluid-structure interactions, many or multi- physical applications, different simulation codes can be linked toger in order to deliver a solution to whole coupled problem. --art method in field f fluid Fraunher environment independent SCAI's code coupling MpCCI provides an interface between most widely used specialized simulation codes. MpCCI takes care data interpolation, synchronizes two simulation codes exchanges data via fast socket communications. Different examples from automotive applications will be presented: simulation f fluid-structure interactions is especially important for design wings spoilers for racing cars, e. g. in Formula 1. The very big CFD model sizes with more than 100 million cellss can be hled by MpCCI whole coupled simulation can be integrated into a batch system work flow. Wheel spoilers every day vehicles are usually made very deformablee plastic componentss distort tremendously while driving. A coupled simulation can be used to find f optimal geometrical material layout for see wheel spoilers. The rmal management cars is anor important topic: MpCCI can be used to couple a shell model thin underhood parts in RadTherm - a specialised widely- used code for radiation, conduction convection - with a stard CFD model flow around a e car. Using full car models forr CFDD radiation simulation very v accurate resultss can be produced. To simulate interaction between air flow around a vehicle chassis suspension, CFDD codes can be coupled to Multi-Body- -System Codes. An nteresting application is a truck exposed to a strong side wind, e..g. while overtaking anor a vehicle or crossing a bridge. During developmentt f- with road vehicles a similar coupling water instead air can be employed to model a car driving through deep water. 6

7 Ulrich H. Rist Institute Aero- dynamics Gas Dynamics Universityy Stuttgart Stuttgart On Need for Expertise when Using CFD - a Teacher ss View University teaching startss with mamatical physical basics engineering, i.e., principles necessary fundamentf tal equations. At a higher level, numerical methods more application specific problems (like boundary-layer flows, for instance) are presented. The primary idea behind this approach is conveying background knowledge not only to future users stware packages but also to future deciders. Knowing how a method works is essential to underst it to foresee where it is applicable where it might fail. The latter hopefully prevents costly pitfalls. In addition, expertise is always essential to judge qualitative meaningfulness new computational results. only contribute to e first partt this process. Some examples related to turbulence modeling, prediction laminar-turbulent transition boundary layer separation re- attachment will be used to illustrate author s view v that expertise is indispensable. A user s expertise will not deliver perfect solution by itself butt it will guide him to look more carefully at certain aspectss force him to carefully study influence critical parameters which anor user (without that knowledge) would have overlooked. In present author s view best way for building up this kind expertise is excellent education followed by a year-long learning process using specific stware packages. The university can 7

8 Pr. Gabriel Wittum, G-CSC, Universityy Frankfurt Below Glossy Surface: Methods, Errors Complexity in CFD Modern CFD stware fers great Graphical User Interfaces (GUI) visualization capabilites. Independent stware vendors (ISV) invested a lot into user interfaces. Easy to use GUIs convincing visualization are decisive arguments for ISVs to persuade customer into bying stware. Below glossy surface, however, core numerics are doing computational work, which is critical part a simulation. This part is about methods, errors complexity finally deciding on reliability computational efficiency results. Often, this does not get attention it actually merits to have, since a lot expertise is necessary to judge this part story. In presentation, we explain interaction methods, errors complexity, i.e. critical part simulation. We furr introduce an approach, how to reach reliability efficiency on large-scale computations.. 8