'Ground Effect - floor' simulation project by Maciek


#1

I created a new simulation project called 'Ground Effect - floor':

A short investigation into diffuser's shape influence on downforce level.


More of my public projects can be found here.


#2

PART 1 | variable length | variable angle

In this topic we move a step further and we’ll try to get to know how changes in diffuser’s basic parameters: length and angle influence its downforce level.

At first I decided to leave both variables unconstrained. Instead I kept upper surface plain and straight. The aim was to have some constant in the system. Now I see I could go even further and do a bit different leading edge, but fortunately, given percentage analysis, it’s not a problem.

The leading edge itself was cut into half and the floor in our “experiment” is highlighted in the picture below.

Next, this time I’ve added boundary layer at the bottom wall (ground). Previously I skipped it as I wanted to focus on the wing only and moving walls were there just to limit the space between the wall and the wing. Also in this case BL is not necessary. To be precise we should remove it and instead set the boundary condition as free slip wall.

Notice: We need to remember that in such reversed situation still it is a “car” that moves, not the whole domain. It means, for no wind condition, there will be no boundary layer formation on the ground surface (!). – Something that so many people tend to forget.

So, why have I done it? Well, simply to provoke some of you into running your own cases and share the results. We could check an influence of inappropriately set boundary condition into final outcomes :slight_smile:

In this case, does it have any influence on our conclusion? No, because I’ve run all the cases with the same settings and what we are interested here are percentages differences not shear values.

Velocity was again set at 40 [m/s].

Ok, here we go with the table:

As we can see the overall drag (this time I’ve included both: pressure and shear ones) increases up to about 10%. At the same time, gains we can obtain properly forming our diffuser can exceed 40%!

(For this case, I assume simulation regular error at something like 2%.)

Now it’s time for two plots:


And finally I’d like to present side views showing velocity fields with stream lines for particular cases:
3 [deg]

4 [deg]

5 [deg]

6 [deg]

7 [deg]

8 [deg]

9 [deg]

10 [deg]

12 [deg]

15 [deg]

What is interesting here? Well, please take a look at the table or plots once more and then run through the pictures. Have you noticed where the maximum downforce value (or equivalently downforce coefficient) appear? What’s characteristic there?

Let’s start from the first picture (3 [deg]), which is automatically our reference point. The yellow colour surrounding the floor refers to undisturbed flow velocity (40 [m/s]). Any hotter one means higher local velocities; any colder indicates lower local velocity values – standard colour pattern. Now, in the second picture (4 [deg]) we see more red under the floor – it means higher velocity, lower pressure, more downforce. This is due to bigger flat plate section in the bottom part. In the next few steps we see even more flat-red and further improvement in results. However, at one point – somewhere between 8 and 9 degrees – downforce values start going down. If we associate this plot region with proper pictures we can spot that this is the moment when first clearly visible flow detachment appears. And as the detachment develops the downforce decreases.

Originally I planned to run the last case at 20 [deg], but unfortunately I wasn’t able to achieve satisfying convergence.

Above: streamlines behind 20 [deg] diffuser.

And lastly, there was one more thing I couldn’t cope with: mesh at the trailing edge. I spent quite a lot of time on this but ultimately I decided to leave it as it is. If you take a closer look there is some influence on flow field, but I think it can still be acceptable. (Case presented below: 15 [deg].)


PART 2 | constant length | variable angle

Ok, here we go with the second part of our investigation. This time diffuser had constant length at 0.35 [m] and the variable was its angle of inclination. I kept the same flow velocity and angle range as in the previous version.

Firstly, let’s take a look at the table and plot.

Downforce plot:

(Downforce coefficient looks exactly the same.)

As we can see this time, in general, downforce goes up pretty much constantly. It’s not straight line, but the trend is clearly visible. And to illustrate the suction effect caused by the diffuser I plotted velocity curve along the domain for two selected cases:

  • 3 [deg] diffuser – the beginning of our investigation
  • 10 [deg] diffuser – the highest downforce level achieved during simulations

The line visible in the upper part of the pictures below was set in the middle of the gap between the floor and the ground. (For reminder the clearance was 0.05 [m].)

3 [deg]

10 [deg]

Finally, longitudinal streamlines (side view) comparison for all analysed cases:

3[deg]

4[deg]

5[deg]

6[deg]

7[deg]

8[deg]

9[deg]

10[deg]

12[deg]

15[deg]

In terms of top surface: I tried to keep it neutral and avoid induced drag generated by transom at the trailing edge.


#3

Impressive @Maciek! :smile: