Undertray Design

[Mikk]

Hey all, I am doing some undertray design work and I am wondering if any of you have done CFD on diffuser angles for undertray tunnels. I know that if I go above 8-10 degrees in the diffuser I will get seperation and I may be able to hit 15 degrees with vortex generators.

Have any of you proven angles higher than this for your diffusers?
Has anyone made successful vortex generators and do you have suggestions for simple ones?

Thanks for any advice/input!

[Luke Phersson]

Hey Mikk,

There are a lot of myths about diffuser design so be careful what you believe. Probably a good idea to have a look at some literature on diffusers, Zhang has published some good work over the years, check out “Influence of Diffuser Angle on a Bluff Body in Ground Effect”. It’s just a diffuser in isolation so not very representative for use on a car with wings/wheels/anything that’s going to be shedding wakes, but it’s still a good idea to figure out the basic mechanisms on how they work.

Don’t get too stuck into trying to optimise the diffusion angle, particularly if you think you can do that with a simple 2D simulation – it’s all about full car interactions and managing the wakes and vortices shed from various components (so 3D CFD with car body, wheels and wings is essential). A great technique is to do total pressure (basically flow energy) sweeps (YZ plane) along the length of the vehicle to see where the ‘clean’ flow can be picked up for the underbody, it’ll also show you how different wakes travel/interact around the car. Given the extremely short wheelbase of FSAE cars and the high CL front wings we use, you will probably find that a lot of your diffuser flow comes from the sides of the car, so beware seals and low ground clearances (counter to most of the literature). Then just iterate, iterate, iterate, make one change at a time, study your results carefully and remember, when it comes to aerodynamics, common sense doesn’t work.

Vortex generators themselves are pretty simple - just a plate with an angle to the flow will get a pressure differential either side, causing a vortex to roll up (Have a look at Journal of Fluids Engineering, they have a lot of good stuff on VG’s). With a bit of practise you can learn to control which way a vortex spins, how to power up existing vortices, how to direct them where you want and how to get rid of them when you’re finished using them. Contrary to popular belief, vortices are not always a bad thing, the more you learn to use them in your favour the better off you will be. Remember that they are simply localised high speed flows, so when applied to an overhanging surface (wing or foot plate underside, underbody etc) they will make you downforce. The tough bit is figuring out why and where you want these vortices, getting them to do what you want, and then stabilising them so that they are not sensitive to yaw, pitch and ride variations. Most of the time we manage our vortices in a reactive fashion, ie we do things with what is already there and can’t be removed (ie front wing tip/endplate vorticies), rather than adding a bunch of new ones to deal with. Keep it as simple as possible.

At the end of the day you’ll need to perform some form of physical testing to validate your designs, after all, CFD stands for Counterfeit, Fraud, and Deceit right? We’ll be pressure tapping our undertray around midyear to see what really happens on track.

Good luck with it,
Luke

[Ally_UGR]

At the end of the day you’ll need to perform some form of physical testing to validate your designs, after all, CFD stands for Counterfeit, Fraud, and Deceit right? We’ll be pressure tapping our undertray around midyear to see what really happens on track.

Good luck with it,
Luke

Can't agree more with this. People at undergrad and even postgrad level throw CFD around like it's the gospel truth. Teams like to say "Oh yeah we used CFD on this and that and our results were brilliant" in the hope that someone will be impressed by it. You absolutely MUST conduct physical testing to validate your claims or *more likely* invalidate them.

CFD software is only as good as the sum of it's parts and the organic being behind the screen. If your model is poor, the mesh incorrect, and your physical data is wrong then you're going to get absolutely nothing useful out of it.

In my own experience, unless you're either incredibly good, have lots of free time, are mad, or get incredibly lucky then playing around with vortex generators should be left well alone. As Luke says, when you know what you're doing they are incredible, Airbus use VG's to great effect as effective wing fences for regulating transonic flow, but they are dealing with a relatively simple aerofoil in clean air. Sticking ones on a car producing sources of turbulence from the front end would require a lot of work, and probably a good few weeks in a flow visualisation tunnel.

Good luck though, just make sure if you DO run VG's and a diffuser to physically test.

[Mikk]

Thanks Luke!

I recently found that article by Zhang and it was very helpful even if it just the diffuser...

And it's nice to know that not everyone trusts the CFD... Too many people take it as true nowadays

[ZAMR]

Originally posted by LukeAUS:

With a bit of practise you can learn to control which way a vortex spins, how to power up existing vortices, how to direct them where you want and how to get rid of them when you’re finished using them. Contrary to popular belief, vortices are not always a bad thing, the more you learn to use them in your favour the better off you will be. Remember that they are simply localised high speed flows, so when applied to an overhanging surface (wing or foot plate underside, underbody etc) they will make you downforce. The tough bit is figuring out why and where you want these vortices, getting them to do what you want, and then stabilising them so that they are not sensitive to yaw, pitch and ride variations. Most of the time we manage our vortices in a reactive fashion, ie we do things with what is already there and can’t be removed (ie front wing tip/endplate vorticies), rather than adding a bunch of new ones to deal with. Keep it as simple as possible.



Luke,

I quite enjoyed this posting, seems like you guys are really going the distance on aero validation. You mentioned capturing wingtip vortexes, but the ones we see are generally being flicked up well above the reference plane, making them seemingly impossible to capture. A lot of our CFD work goes into finding "clean" air for the cooling ducts and avoiding this vortex. I know you probably don't want to reveal any trade secrets but how have you been able to keep the tray simple yet still find a way to utilize this vortex?

PS I love your guys' recent cars.

-Zach