I am out on a limb here but............

Does anyone know of any free CFD / Virtual wind tunnel software available?.

What I want to do is produce a simple 2D profile, my bike fairing/windscreen and see the wind flow over it, and then I can see what adjustments I can make to alter the windflow.

Any ideas?, or am I pee'ing in the wind here

Use a fan and pee on your bike.

OR... drive really fast and ask someone to ride ahead of you and pee on you-- your bike.

Yeah, that'll work.........

Anything else that doesn't involve urine though?

Use a hosepipe?

Glue wool onto your fairing (sticky tape) in various places, and see how they move when you ride. Lots of wiggling = turbulent airflow = drag. Not much wiggling, but streaming in direction of air = laminar flow = good. Hanging down under it's own weight = no air movement in this area at all = region of drag probably.

And remember, on a bike the aerodynamics change MASSIVELY depending on where the rider sits - tucked in behind the screen, or sitting up, leaning etc. Just reducing drag by optimising the airflow when you're tucked in could mean that at road speeds you have too much/little pressure on your chest, making your arms tireder etc....... Good luck though!

Cheers, I'll take a look at MicroCFD.

@ Tristan, yeah, the problem I have at the moment it this......

Currently the screen on my bike is fine up to about 90mph, then I need to duck down under it to stop the buffeting over 90mph, now this is no biggie but..........

When I go to Germany I am going to be using a tankbag, which means I am not going to be able to get "under" the screen at high speeds, I want to buy a new "double bubble" screen, but wouldn't mind checking out how this will change the airflow before I shell out the £60 for it.

Buy it. Fit it. Cover it in clingfilm so it stays clean. If you like it, keep it (and remove the clingfilm), if you don't then send it back (it'll be clean and returnable with the clingfilm).

CFD is difficult. Making a rough approximation of the model is hard (how much detail; is 2D enough etc), meshing it well can take ages (different meshes can have a HUGE impact upon final results), and wat sort of turbulence model etc you use will a) effect the convergence times, which might be around 10 hours for a decent 2D CFD model of a motorbike and b) effect the accuracy of said results even after 10 hours.

In short, clingfilm and wool is by the far the best method for someone worried about not wasting £60. If you were worried about wasting £200,000 on a racing bike then I would consider CFD.