Optimize Airflow 3D model design , how to?

Discussion in 'Design and Modeling' started by janis, Jun 17, 2014.

  1. janis
    janis New Member

    I did search on this forum but couldn't find a topic about it..

    i made 3 pipe that goes over to 1 pipe out put..

    i wanna optimize the model for airflow. i have made it in Blender.

    but how do i do this? do i need other program?..
    i wanna see the airflow and what parts will heat up if the airflow is hot

  2. AmLachDesigns
    AmLachDesigns Well-Known Member
    I assume you are talking about some kind of engine exhaust system?

    What you are talking about is not easy as it relates to fluid dynamics (yes a gas is a fluid): the equations used to model fluid flow are not simple to solve, so numerical methods are used to get approximate solutions. The software that is used to solve the kind of problem you have is called CFD (Computational Fluid Dynamics) software, and I imagine is not so easy to use (although I did find some open source software online).

    There is someone posting here that maybe able to help as I imagine they are using this kind of software, although your problem is harder because it involves gases (they compress and liquids typically do not) and you are also interested in thermal properties.

    If you can't be bothered with all that, here are my simple ideas to optimise your pipework (assuming you don't want a 'tuned' 2-stroke exhaust):
    - A bigger pipe will always allow the gas to flow better (faster) than a thin one;
    - When you change diameter of the pipe do it as gradually as possible if you are reducing: if you are increasing, it is not so important, but avoid step changes;
    - make bends in pipes as long and as gradual as you can;
    - For the heat aspect, imagine your gas flow is carrying sand or something - where would the sand hit? Those are the places where the pressure will be higher and the heat transfer greatest, I think. So anything that protrudes into the flow, any place where a pipe diameter reduces and the outside of bends.

    Having said all that, I hope your system is for a model engine otherwise it may be too expensive/too big to print.

    Good luck!
  3. janis
    janis New Member
    thank you :)

    owke CFD in 2 months i have time to look into ..

    but isn't there easy/qicker software to do this?

    it's for model. for me the airflow need to optimized. i (hope) don't think there will be compression

    maybe liquid will do the job to at this moment....

    but still how to do that? :)
  4. FreeRangeBrain
    FreeRangeBrain New Member
    There is no quick and easy way to calculate airflow.

    There is NO quick and easy way to calculate airflow.

    Liquid flow is challenging. Airflow, with compressibility of the media thrown in on top, is Nature's way of thumbing its nose at your mere mathematics.

    Per the previous reply, the use long tapered sections for transitions reduces the resistance to flow, but only to a degree. Ø3" to Ø1" is a 9:1 reduction ratio The only way you escape compressing the gas is to be driving the flow with suction on the small end, not pressure on the large end, or by having the gas spontaneously and dramatically cool itself just as it reaches the constriction - possible, but highly unlikely. Any change in cross sectional area or flow direction will cause resistance. Even a pipe exiting into a large tank calculates as a resistance to fluid flow, particularly if the exit is abrupt.

    I would expect that any heat transfer to the pipe will be dependent upon localized air density and the corresponding heat energy density, combined with fluid viscosity, surface friction and surface texture. A compressed gas has the same BTUs as an uncompressed gas of the same mass, but held in a smaller volume, thus a higher energy density, resulting in a higher temperature. Pressure losses occuring in fluid flow usually calculate out as heat rise due to energy conservation laws. Surface friction losses in gas flows are usually insignificant due to the low density of a gas versus a liquid. Losses due to turbulance, however, can be tremendous. An aircraft wing entering a stall condition doesn't just lose lift. It generates large amounts of drag due to work exerted in the generation of turbulence. One wing will usually stall before the other, induce excessive drag on that side of the aircraft, and induce a spin. The roll caused by the loss of lift will often increase the angle of attack, exacerbating the stall.

    I recommend looking up Bernoulli and Reynolds equations as a start. Move on from there to "line losses."
    Last edited: Jun 17, 2014
  5. stop4stuff
    stop4stuff Well-Known Member
    If your 3 into 1 tubing is for an exhaust system, have a look at a 1300 MKII VW Golf exhaust manifold and front downpipe, or going back earlier, a Lotus Cortina exhaust system. CFD is great for optimising flow, but the right software that can give accurate results is a bit of a holy grail and will cost more than your average family car.

    JACANT Well-Known Member
    As all of the replies said. It is not that easy.
    You need to keep all of the walls and transitions as smooth as possible.
    I downloaded the file from the original thread.
    There are a few mistakes in it.
    I have made a copy for you to see what is an approximation of what you need.
    You may find it easier to use this software, Designspark, it's free http://bit.ly/1kEs7Bo
    Use the loft command.

    Attached Files:

  7. AmLachDesigns
    AmLachDesigns Well-Known Member
    Hi Jacant,

    nice job!

    What was your workflow? Did you make it in Blender or import the .stl?
  8. janis
    janis New Member

    Thank you for the file :D

    can you explain how you did it? and make it so smooth?
    because i see a lot of vertex in there..

    in the attachment you see the last version i traid to make (in total 11 hours of work :( lol )

    i saw that my size wasn't great like it should be..

    but is it possible that you can change it a bit...

    the 3 hole's they need to have 0.8mm inner ring
    and the 1 hole in the back end it need 1.05mm inner ring till 0.5mm in the tube then it need to be 1mm


    Attached Files:

    Last edited: Jun 18, 2014
  9. janis
    janis New Member
    thx for the replay... owk CFD and designspark it is then.. just need to find the time to learn it a bit..

    wow that looks great..
  10. JACANT
    JACANT Well-Known Member
    Imported STL from AutoCad.
    I started with the outside and inside circles at the ends of the pipes, placed at the correct positions in the height.
    Copied the top circle three times. Copied and moved the three bottom circles up to the joint you see.
    I then used a 'Loft' command, three times on the top circle and the upper three circles, following the normals to each, which created a smooth transition. (I did try an Array, it didn't seem to work)
    Did the same to the inner circles. Boolean 'Join' the outside circles together, same with the inner circles. Boolean 'Subtract' inner from outer shape.
    Extrude bottom 'Region' to meet top shape Join both together.
    Fillets to inside edges where required.
    It can be done in DesignSpark. (free)
    Last edited: Jun 18, 2014
  11. AmLachDesigns
    AmLachDesigns Well-Known Member
    Cool, thanks - wasn't sure how I would do it in Blender to get such good merging of the tubes. I might have to give it some thought and see if can do it...

    Thanks again
  12. JACANT
    JACANT Well-Known Member
    Import the file into Blender.

    Attached Files: