Announcement

Collapse
No announcement yet.

Airplane contrails

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Airplane contrails

    Hi.

    I'm attempting to simulate airplane contrails in an effective manner. Currently have two simulations running; one using pflow to emit from, and one with just a couple cone-shaped objects.

    Below is kinda the look I'm after, but I'm not picky, just need something that looks fine. We will be seeing it from a medium distance, and from all angles.

    https://gifer.com/en/Wm9U

    And this is the animation it will be put into: https://vimeo.com/321233902 Though will a bit different angles.

    So, the reason for making this thread is to see what suggestions you guys might have, as I'm having issues simulating this in an effective way, meaning that the plane is a Tupolev Tu-160 moving at realistic cruise speed of 960 km/h - which I would think requires a lot of steps per frame! Currently I'm at 80, but no idea if this is enough. Would it be better to cheat this in some way? Basically all I need is smoke being shot out of an emitter at great speeds. Also of course I'd like some help to make it look good, as my current results are not great.

    Attached are some screens of what I have currently and settings, for the non-pflow setup.

    Any help or guidance or pointers or anything! Is very much appreciated!
    Last edited by crimea; 26-04-2019, 04:30 AM.

  • #2
    Hey, definitely a high number of steps per frame would be needed, but also not sure if 80 would be enough. Note that all vfx simulators model fluids using math that is meant for speeds way below supersonic and start to diverge with reality at 200-300 kph, so it's very likely that you would need to cheat it. Off the top of my head - less conservation might be needed so you'd get a more laminar flow - you should not need the high rolling and fluidity that high conservation contributes. Instead, maybe you'd need high classic vorticity or massive vorticity which is surface only and high scale is zeroed. Also, give forward transfer advection a shot - it usually produces interesting results for such fringe cases.

    Hope this helps!
    Svetlin Nikolov, Ex Phoenix team lead

    Comment

    Working...
    X