Difference between scales on wind tunnel

[adam2003]

Whats the difference from using a 60 scale compared to 100?

Cost of parts?
Do you get any differences in how parts work with different scales?

[astracrazy]

100% will obviously be better, but due to the regs its 60%

teams will i guess always go for they biggest they can have

[Greg Locock]

Whats the difference from using a 60 scale compared to 100?

Cost of parts?
Do you get any differences in how parts work with different scales?

I suspect 60% scale parts cost exactly the same or even more than real ones. Not all effects scale in the same way with length, for example I'd guess that size (scale diameter at a given scale length downstream) of a vortex train is proportional to linear scale, yet the pressure it exerts may not vary as L^2. So you end up building two CFD models, one is the 60% version which is calibrated against WT data, and the 100% one which you use the 60% to debug. However i don't work in wind tunnels or CFD, I just use the results.

[DaveW]

Not all effects scale in the same way with length....

"Dimensionless Numbers" are used to estimate the effects of scale. Reynold's number is one, here are a few others.

[Greg Locock]

Yes, but different physical effects have different dimensionless numbers.

for example boundary layer /thickness/ doesn't scale with reynolds. From what I remember very badly boundary layers are too thick in small models.

[aussiegman]

The biggest issue with smaller scales is that the air increases in "apparent viscosity" of the air flow making the apparent flow too "thick". As the Reynolds number is based on a ratio containing the effective viscosity of the fluid (in this case air) it makes for problems in scaling of boundary layer effects and interfaces when compared to the 100% scale.

The surfaces may simply be too small to effectively model the turbulent airflow over the given surfaces or the boundary layer effects as the effective viscosity of the air to the surface has changed due to scale which impacts elements such as drag, eddy and vortex formation etc. This causes problems with the normal "cascade" effects seen between inertial/kinetic forces and viscous forces which are determined by a the normal dynamic viscosity coefficient for air which is altered by scale of the model which changes the apparent surface area etc.

For flowing testing you would typically use the largest possible scale available to reduce error and influence from scale differentiation.

[DaveW]

for example boundary layer /thickness/ doesn't scale with reynolds. From what I remember very badly boundary layers are too thick in small models.

Absolutely, but Reynolds number also suggests what is required to minimize that particular effect. Other affects of scale require different numbers to be matched (e.g. Strouhal number for flow instabilities - might be important if vortices are used as virtual skirts).

Greg & assieman illustrate well the problems of scale models, & why full scale testing is generally the preferred option.

[hardingfv32]

I would propose that the issues that are created by the scale models are out weighed by the issues that a tunnel test section for 100% models creates. The teams had an opportunity to develop 100% tunnels before the tunnel regs were implemented and none did so.

Brian

[wesley123]

I would propose that the issues that are created by the scale models are out weighed by the issues that a tunnel test section for 100% models creates. The teams had an opportunity to develop 100% tunnels before the tunnel regs were implemented and none did so.

Brian

iic a few teams did. I'm sure Sauber has a 100% windtunnel, and I beleive the same counts for McLaren and Mercedes

[WhiteBlue]

I'm sure Sauber has a 100% windtunnel, and I beleive the same counts for McLaren and Mercedes

You can run every tunnel at higher scale if you reduce the speed. But that is not meaningful for F1. The Sauber 100% capability was primarily used by Audi for road car testing. Those studies do not go much beyond 200-250 kph.
Sauber's tunnel is powered by a 3 MW motor. The bigger the beast the more reserves you have. I don't know if there are higher powered tunnels now in the industry but back in the time when BMW purchased the team it was state of the art.

[marcush.]

I would propose that the issues that are created by the scale models are out weighed by the issues that a tunnel test section for 100% models creates. The teams had an opportunity to develop 100% tunnels before the tunnel regs were implemented and none did so.

Brian

i beg to differ :

The Mercedes tunnel -formerly Honda Racing-in Brackley has 100% capability ,as has Windshear in USA , I think one of the Toyota Tunnels ,Sauber of course and some more.

Why on earth MGP has only now opted to run 60% models is beyond me but I would think it is very hard for a tyre company to develop two different size windtunnel models for the teams .Inevitably one or the other will have more difficulties to to be matched to reality ....It looks like the teams using 60% do fare better ,but then they are the ones who have the biggest budgets..did Bridgestone supply "real"scale model tyres as well ?

One would think the biggest issue with smaller scales is to accomodate all the phsical effects you want to make use of in full size later on:
how do you reproduce the exhaust plume ,and how will vortex shedding translate to reality bearing in mind your model will hardly show the same aeroelasticity effects than the real thing ...not even speaking of dimensional accuracy issues with smaller scales ..

[langwadt]

I would propose that the issues that are created by the scale models are out weighed by the issues that a tunnel test section for 100% models creates. The teams had an opportunity to develop 100% tunnels before the tunnel regs were implemented and none did so.

Brian

i beg to differ :

The Mercedes tunnel -formerly Honda Racing-in Brackley has 100% capability ,as has Windshear in USA , I think one of the Toyota Tunnels ,Sauber of course and some more.

Why on earth MGP has only now opted to run 60% models is beyond me but I would think it is very hard for a tyre company to develop two different size windtunnel models for the teams .Inevitably one or the other will have more difficulties to to be matched to reality ....It looks like the teams using 60% do fare better ,but then they are the ones who have the biggest budgets..did Bridgestone supply "real"scale model tyres as well ?

One would think the biggest issue with smaller scales is to accomodate all the phsical effects you want to make use of in full size later on:
how do you reproduce the exhaust plume ,and how will vortex shedding translate to reality bearing in mind your model will hardly show the same aeroelasticity effects than the real thing ...not even speaking of dimensional accuracy issues with smaller scales ..

I don't see how the tires matters much other being the right shape and size

I wonder if they have tried, or do, run in partial vacuum or some mixed gas to "scale" the air aswell?

-Lasse

[Forza]

Toyota TMG has two wind tunnels: WT1 for full scale and 60% model and WT2 for 60% model testing. Fan power is 2.3 MW.

[PlatinumZealot]

Whats the difference from using a 60 scale compared to 100?

Cost of parts?
Do you get any differences in how parts work with different scales?

Reynold's numbers will be different and that changes the behavior of the fluid. Mathematically though it is the ratio of inertial forces to viscous forces. Physically speaking it is characterized by the relative size of a small feature in the flow such as a vortex, compared to the size of the volume of interest. For example swirling a spoon in a tea-cup has a larger reynold's number than swirling a spoon in the ocean.

If you translate that to an F-1 scaled model: the smaller model will be more affected by tiny disturbances in the flow, hence it will be inaccurate and that will throw off your analysis.

[riff_raff]

The biggest issue with smaller scales is that the air increases in "apparent viscosity" of the air flow making the apparent flow too "thick". As the Reynolds number is based on a ratio containing the effective viscosity of the fluid (in this case air) it makes for problems in scaling of boundary layer effects and interfaces when compared to the 100% scale.

The surfaces may simply be too small to effectively model the turbulent airflow over the given surfaces or the boundary layer effects as the effective viscosity of the air to the surface has changed due to scale which impacts elements such as drag, eddy and vortex formation etc. This causes problems with the normal "cascade" effects seen between inertial/kinetic forces and viscous forces which are determined by a the normal dynamic viscosity coefficient for air which is altered by scale of the model which changes the apparent surface area etc. For flowing testing you would typically use the largest possible scale available to reduce error and influence from scale differentiation.

aussiegman- I'm a mechanical guy and I know next to nothing about aero, but I did spend some time working for a company that built wind tunnel models for the aerospace industry. Many of the models we built were of very reduced scale since they had to fit into the small area test section of "blow-down" transonic/supersonic tunnels. The smaller the model scale, the tighter the profile tolerances became that we were asked to hold. Regarding Reynolds number and scaling effects, I recall that one or two of the more sophisticated tunnels had the ability of using a different mixture of gases (instead of compressed air) that gave more accurate test result, but I don't recall what these gas mixtures were composed of.