I know the best way to get the best analysis is to use full mockup, but what is the smallest scale you can use and get the best approximation of aerodynamic efficiency in a car?
I don't know too much about Reynolds numbers but they usually appear in a scale. I forgot the scale but these numbers usually determine the type of flow: laminar, transitional, or turbulent. There are a bunch of equations calculating Reynolds numbers, and you take measurements from the tests to calculate Reynolds numbers.
Say you get a figure at 40% scale that might be a turbulent Reynolds number; at bigger scales like 50% and 100% the numbers don't match exponentially and you might get a number, at these higher scales, of laminar flow. If that makes sense. So usually you want to run bigger scales to get more accurate results.
Bring back wider rear wings, V10s, and tobacco advertisements
For a meaningful simulation the requirement is the fluid dynamics similitude (it comes from the non-dimensional form of the Navier-Stokes equations) : Reynolds number and Mach number of the windtunnel test must be equal to the Re and Ma of the real phenomenon. (actually there’s also the Froude number but it’s meaningful only in few situations)
The definitions :
Reynolds = (density * velocity * characteristic length) / (dynamic viscosity) = velocity * length / (cinematic viscosity)
Mach = velocity / sound velocity
Sound velocity and cinematic viscosity depends by temp of fluid.
Now assume the real phenomenon is a car running at 200 km/h (55 m/s) in straight line and we have a 50% model to test. To have the same Re you have to use, in the wind tunnel, assuming air characteristics (density, viscosity and temp) are the same, twice the original speed => 110 m/s. Looking at the Mach, since the temp is the same, you have twice the Mach, if speed of sound is 340 m/s (happens with temp = 15°C), the real Ma is 0.16 and the simulation Ma is 0.33. Fortunately at low speed that’s not a big problem because the important thing is to be in the same range of compressibility effects and up to Mach about 0.3 the compressibility effects are basically negligible. Anyway it would probably be better to increase a bit the scale, or to reduce the speed of the real phenomenon to test because in this particular case we are at about the limit of Ma. So, as you see, to simulate a real speed of 200 km/h, you need at the least 50% scale, to test a real phenomenon with higher speed you need a bigger model. Then you have to consider that the energy consumption to move the fluid depends by both the dimensions and the speed of air in the tunnel.
Another possibility is to modify the fluid characteristics, as I said the cinematic viscosity depends by temp (for air it increases with temp) so a cooled wind tunnel allows to use smaller scale or higher speeds with the same scale still keeping the same range of Ma (that’s because the speed of sound reduction reducing the temp is slower). But obviously a cooled wind tunnel costs a lot.
Then you have always to consider the model (that costs A LOT) the tolerances and the structural characteristics + the fact that inside it you need balances and instruments would lead to a bigger model, etc...
looks very much like a tunnel is a tunneland you only know what you get if you really have a possibility to crosscheck your findings and have lots of experience with that particular tunnel.
Ad to this the fact that this all costs a fortune to run the results just have to be good and accurate and we have the reason for the tunnelhype not translating into laptimes.
I feel that if you go for general shapes a rather small tunnel will do but if you try to develop the details wich are of major importance nothing beats
the real thing.
What i don´t understand is why going thru the hassle and build a scale model wich needs to be very very accurate and stiff to a scale of almost the real thing instead of putting the real car into a tunnel and correct for the blocking instead.This way around you have just one source of fault,whereas in scale the errors seem potentially endless....
Sauber does lots of fullscale testing now is this correct?
with the advent of very affordable supercomputers (AMD Opterons, Yeah!!), how accurate or how close can a computer simulation come to the real testing results when it comes to aerodynamic analysis?
computers are still dependend on what the idiot sitting before it is feeding in.So the bullshit numbers just get crunched faster.
seriously,if your models are correct the simulations will work phantastically.In case your sims do not capture what really is happening it´s a waste of time.
so, theoretically, you can design the car, test it and try different aerodynamic solutions on the car while not getting your hands dirty? has anyone done this? Sounds like a good cost cutting measure to me... simulating everything on the computer... your thoughts?
(I know that Commanche helicopter was designed and tested on a computer and then first prototype was produced...)
assuming your simulations capture all relevant aspects correctly ,yes of course.
I do have some very basic engine simulations to feed my setup programmes.
It is always funny to see the computed graphs against the real world curves...with engines I feel you pretty much can simulate what will happen in reality.And you can look for the sweet spots of this particular layout also...it works.
Of course a real engine expert would do things right anyway.And the programme depends on the parameters influences being correctly weighted .
so, does anybody know how much time top teams (Ferriri, Renault, etc) spends on computer simulated aerodynamic analysis as opposed to real wind tunnel time?
Computational Fluid dynamic is a highly powerful tool, and depending on the type of flow to be analysed, the software can be more useful than wind tunnel testing.
I quote: "Computational fluid dynamics has already had a major impact on airplane design, and a recent prediction from the national research council calls for CFD to become the critical technology for aerodynamic design over the next decade.
Historically, the early development of CFD in the 1960s and 1970s was driven by the needs of the aerospace community. However, modern CFD cuts across ALL disciplines where the flow of a fluid is important"
The advantage of CFD is the all the data is there after the solution is procesessed. You have instant access to every detail of the flow. Which wind tunnel need complex equipment to gather such data.
Yes, the CFD analysis is dependent on assumptions, as direct analysis of every single molecule of air (which would give 100% accuracy) is far far beyond reach for our technology. Assumptions are needed to ensure all properties of the equations are fullfilled. Although this is the case, very very accurate results can be obtained provided the CFD user is a confident and skilled user that understand the mathematical background of the CFD code. providing the properties on the flow are as accurate as we can input, then solution becomes dependent on the mesh, as i mentioned, for a pure accurate solution the limit must be 0 (direct molecular analysis), but this is not possible, so it all depend how close to the limit we can achieve...........deopendent on computing and time costs.
Russell Harrison
So, yes, a computer system
Russell Harrison
Forced Convection Design Engineer, Comair Rotron Europe Ltd
CFD is based around assumptions; the accuracy of the solution depends not only on the knowledge of the mathematics behind the software but the assumptions the user makes!!!
You will probably find......correct me if Im wrong, that the bigger teams will have seperate departments within the aerodyanamics department, that will gather data from both CFD and wind tunnel, and corrolate
Russell Harrison
Forced Convection Design Engineer, Comair Rotron Europe Ltd
CFD is based around assumptions; the accuracy of the solution depends not only on the knowledge of the mathematics behind the software but the assumptions the user makes!!!
yes, it seems as tho Ferrari has a state of the art computer analysis system at their wind tunnel facility.. it would seem logical that they would employ CFD in thier arsenal.
yes, it seems as tho Ferrari has a state of the art computer analysis system at their wind tunnel facility.. it would seem logical that they would employ CFD in thier arsenal.
Of course, my line of questioning has a point...
can any of you guys envision a time where wind tunnel testing becomes obsolete?
The major design concept that millions and millions of pounds is being poured in to is the development of a hypersonic transatmospheric vehcile.
Unfortunately, no ground test facilities/wind tunnels exist in all the flight regimes covered by such higher mach numbers and high flowfield temperatures to be encounted by such vehicles, and the prospects for such wind tunnels in the future are not very encouraging. Hence, the major player in the design of such vehicles is CFD.
Russell Harrison
Forced Convection Design Engineer, Comair Rotron Europe Ltd
CFD is based around assumptions; the accuracy of the solution depends not only on the knowledge of the mathematics behind the software but the assumptions the user makes!!!
