Vespa CFD/3D Model

From $1

    Project Description

    This is a project started as part of the Dry Lake Racer: CFD (Computational Fluid Dynamics). For that User:internetscooter (a.k.a. Paul McIntosh) paid to get his scooter 3D scanned through scan-xpress.com.au in Melbourne. For this project the data is being provided free of charge under the condition that no money be made from derivative works unless, the derivative work is feed back into Vespa Labs for the benefit of all.

    Source code and case settings are available at: https://github.com/internetscooter/Vespa-Labs

    dry_lake_racer.png

    Information from the original CFD (Computational Fluid Dynamics) lab notes will be uploaded and one of the Vespa 3D Model datasets suitable for CFD analysis. The complete dataset contains higher resolution detail (e.g. below) as well as scans of more standard Vespa parts. Please contact paul mcintosh at internetscooter dot com if you wish to experiment on that data. This project and data would suit university students wishing to use real world data for CFD or other modelling projects.

    vespa_full_scan.png

    The data is available for registered Vespa Labs users here: Vespa 3D Model

    CFD Modelling

    Here is an example of what can be done with the 3D Model using open source CFD tools like OpenFOAM. As the project progresses there will be OpenFOAM examples that people can download and tweak themselves.

    Get Adobe Flash player

    The above shows a test run with a "real" Vespa model but a fake rider. This shows things like the front indicators contributing a lot to the aerodynamic drag, the big red areas indicate high negative pressure zones. The above was done as a test on the VPAC and MASSIVE HPC (high performance computer) clusters. Students and researchers in Victoria, Australia have access to these clusters and may wish to consider HPC projects using the Vespa data. We encourage other HPC researchers to contribute their workflows back to Vespa Labs (or provide links to publications and cite where the data came from via the url).

    Research Questions

    Note that the model data also includes more standard Vespa parts (rather than the race adapted version shown). Here are some research questions that would be good to answer using this data:

    1. What is the aerodyamic drag of a Vespa?
    2. What is the lift force on the front of a Vespa at various speeds?
    3. How much does a windscreen reduce drag?
    4. What is an optimal windscreen design?
    5. What is the optimal windscreen design for reduced helmet noise?
    6. What is the drag impact of various sitting positions (regardless of windscreen)?
    7. What are the areas on a Vespa that cause the most drag and what are the solutions for reducing drag in those areas?

    See also: Aerodynamic Investigation of a Scooter in the University of Perugia Wind Tunnel Facility by Stefano Ubertini - Univ. of Rome Tor Vergata and Umberto Desideri - Univ. of Perugia.

    And review (and update as needed) Aerodynamics

    What's Important

    Here are some bullet points on what areas of CFD to focus on for the problem of Vespa CFD (so you don't waste effort learning stuff that is not applicable)

    • incompressable flow: we aren't going "that" fast so we don't need to model air compressing as part of the problem so we can just use incompressible flow models
    • bluff body: we are modelling something that is not streamlined. An aircraft it streamlined so we want to know about what is happening at the surface (friction drag) but a Vespa stirs the air up so we want to know about how this causes drag (pressure drag).
    • Reynolds-averaged Navier–Stokes (RANS): equations are (at the moment) the most used in this style of CFD
    • Turbulence Models: are required to work out the turbulence part (since we are a bluff body this is important) e.g. RANS based turbulent models and OpenFOAM list
    • Large Eddie Simulation (LES): give a better resolution of flow (requiring more processing power). LES takes the approach that the large eddies influence things a lot and can be modelled by themselves while the small eddies can be ignored and dealt with via RANS. Fortunately OpenFOAM comes with both RANS and RANS+LES motorbike examples.
    • The state of the art (2013) is that CFD can predict drag +/-5%, however we need a resolution of <0.5% to do useful measurements (i.e. make small changes and say they improve Cd) - so CFD is more about predicting/analysing where problems are so that further tests can be caried out (i.e. wind tunnel or real life riding).

    Additional Resources

    Here are some resources that will help anyone that is learning CFD from scratch...

    CFD seems to be typically taught by building on a lot of background learning, so by the time you are learning CFD at Uni you are just piecing together a number of things that you already know. If you want to learn backwards where CFD knowledge is your goal and you haven't done supporting training, then you don't know what you don't know. CFD books assume prior knowledge and won't even give a glossary of common terms that you could Google. Here is some info I have used to get some understanding. These are listed in order of required learning but you can also start at the bottom and search back up the list to find the missing bits you need as you need them.

    Some other links that I have found along the way...

    Books

    Tags: (Edit tags)
    • No tags
    FileSizeDateAttached by 
     book.jpg
    No description
    10.21 kB20:10, 18 May 2013internetscooterActions
     dry_lake_racer.png
    No description
    241.99 kB23:01, 19 Aug 2011internetscooterActions
     vespa_full_scan.png
    Example of full res - front legshield logo
    271.94 kB18:07, 20 Aug 2011internetscooterActions
    Comments (2)
    Viewing 2 of 2 comments: view all
    Some Resources I've collated for learning CFD : http://cfdrevolutions.weebly.com/cfd-course-material.html
    Posted 21:23, 17 Aug 2012
    thanks - i added the link to the list
    Posted 21:37, 17 Aug 2012
    Viewing 2 of 2 comments: view all
    You must login to post a comment.

     
    Powered by MindTouch Core

    Disclaimer: Vespa Labs contains information that is VERY likely to wreck your scooter and possibly yourself both intentionally (i.e. gaining more peformance while sacrificing reliability + safety) and unintentionally (i.e. misleading or incorrect information). Vespa Labs is only a wiki and intended as a dumping ground for information and not as a properly reviewed source. The same disclaimers that wikipedia.org use apply to Vespa Labs. The short version is use information at your own risk, both information on the main wiki and in the user areas are intended to be used only as "thought provoking" for someone that knows what they are doing. If you try to implement a "thought" Vespa Labs is not reponsible and if something goes wrong we hope that you live and update the offending information with corrections to warn others.

    Vespa Labs is an international site and therefore may contain information that is not road legal in some countries/states and may also invalidate insurance policies. Treat all information as experimental and for "race use only" (i.e. not for road use - even if it is implied or explicitly stated). Refer to and adhere to your local road and legal rules, as well as the manufacturers recommendations.