Aerospace Engineering hypercar thesis, part 1

After my previous project there finally comes the time to write my undergraduate thesis. Somewhere around three months ago, I was pouring over the details of Aston Martin Valkyrie, trying to best understand its design and reason for every shape's shape. While I am, admittedly, an aerospace engineering student, I am also a manqué automotive design student. I was really serious about this during my middle school and even got in touch with a very supportive professional designer, but with a weak English and lack of parent support the dream withered with time. I still have a high interest in design and follow many designers on Instagram. A project that combined aerodynamic (my other interest) and aesthetic design would be a both amazingly interesting and maybe a way of finding a job in the industry, as well as a symbolic middle finger to life. You can't stop me.

From left to right: Adrian Newey's initial Valkyrie drawing, my whiteboard sketch, my paper drawing and finally the NX model at the current stage.

Initial design

From the beginning I wanted to follow Adrian Newey's Aston Martin Valkyrie. This car is in my opinion a completely new paradigm of performance car design and to learn it best, I have to design something similar from scratch and run a bunch of Fluent calculations. I started from Adrian Newey's drawing of Valkyrie layout from his book.

Initial design drawing of Aston Martin Valkyrie.
Photo of Adrian Newey's book "How to build a car"

While I am not designing an entire car completely with suspension, powertrain and all that, I wanted my design to be somewhat grounded in reality. So what I've done was print the standard 2D human manikin, cut it and shaped in a way I wanted a driver to seat. Then I added the entire car to scale and then copied the drawing onto a A1 sheet of paper. The tub is (I think) reasonably large, there should be enough space for powertrain etc.

Then, I showed it to my advisor... and he didn't like it very much. There were many reasons for that, but we finally agreed to surround the tub with huge diffuser-like tunnels on each side instead of weird splitter-wing solution I guessed. Later on I realized that Valkyrie's front wing and negative space works exactly the same way - the wing turns the air upward with encounters increasing amount of space down the flow acting like a diffuser. I guess our way is just a bit less elegant.

My paper drawing with scribbles added during a discussion with my advisor.

Because we had some issues communicating only with our pens and paper, my advisor asked me to start working on the initial 3D model. Since my faculty has a license for Siemens NX and we learn it during classes, I decided it would be the best software to go. Soon I realized that modelling geometric shapes is one thing, and those organic shapes I will have to implement here is a completely different ball game. Even though during the advanced CAD classes I finished tasks half an hour before everybody else, this time I struggled immensely.

Initial assembly in NX.You can see the front right wheel turned left and right to determine how wide can tunnels be.

At this point it is easily visible that the tunnels are laughably narrow compared to their height. I'm not sure if that would work, but it definitely looks bad and even worse, doesn't look like efficient utilization of all that space we saved with that narrow and raised tub. What became evident after transition to 3D is that this blocky tub created by extruding the drawing is not the most efficient way we can do it.

Left: Aston Matrin Valkyrie. Notice the flat V-shaped tub.
Right: my sketch of the tub sections before I could model it in NX.

If the tub is shaped in the more organic shape, with the lower edge blended just to leave enough space for the occupants, the tunnels could be either widened and reshaped to utilize some extra space (and lose that rectangular cross section), or also be slightly rotated towards the center and also reshaped and widened. The latter would decrease the negative space under the raised tub though, and I wanted to use it to push some air below the car into the rear tunnels.

After another discussion with my advisor, the next point was to reshape the front tunnels such that they are wider and a bit lower at the front - what is important is the cross section area that should increase linearly. The idea about pushing the air downward to the rear tunnels through the center opening was accepted, but the nose was way too high and sharp. Also since we expect a huge downforce on the rear of the car from larger rear tunnels and rear wing, we need to add some downforce on to balance everything out. shaping nose to get some pressure buildup seemed like a good idea.

Later I had some exams and a family skiing trip, so my learning of surface modelling in NX was somewhat constricted. I had a few bits of time to draw though, so I got my old Wacom tablet and did a quick rendering over the NX screenshot you see above.

In a week I learned a bit more about NX and so here is the updated model:

Much better, isn't it?

Aesthetic design

I know that logically, this shouldn't be even a slightest concern to me at this point. This is a road car though, and also there is plenty of moments during my week when I can sketch and cannot work on NX. Also both Marek Reichman from Aston Martin and Adrian Newey from Red Bull said that aesthetic "upper" and aerodynamic "lower" parts of the car were designed rather independently.
The first "research" sketches were mostly consistent with my initial aerodynamic ideas, but the main theme stays the same - I want the cleanest, most elegant shapes possible, only constrained by aerodynamic and structural design and more geometric and industrial-design-like approach where the aesthetics and engineering meet.

Some of my initial sketches from my notebook combined with some older sketches I did around a similar idea.

I haven't been drawing much before this project started so I am quite rusty. Expect to see the biggest progress in this section in the next project update.

Notes on the NX Workflow

My university's advanced CAD course was indeed more advanced than the basic one, but definitely not sufficient for something this big and "curvy". Therefore I had to come up with smarter ways to plan the whole project. This is probably going to be laughable for any professional with experience in large projects, but could be useful for people without experience (like me when I started this).

1. Order of modeling. While in the later stages I am definitely going to edit each component in the assembly in various order, it was useful to determine a priority and dependence based order of modeling parts. I started with the tub and added wheels in the inital position, then after that front tunnels and finally mock powertrain and rear tunnels (these two I messed up). Once these components are done (and probably redone a few times), I am going to start the bodywork.
2. Sensible relying on drawings. This is my first project where I have to really use Studio Splines, and boy they are unruly. I had a very hard time trying to get the required shape without overconstraining the curve.
3. Clear, highly linked and modifiable sequence of features. I wanted to be able to correct those complex surface-modeling-based shapes with minimal effort. This is why my files always include plenty of datum planes based on another planes, intersection points with curves and surfaces and so on.

Truth to be told, I must have messed something up, because at this point the powertrain component kind of took over tub and other parts in the assembly and I have no idea how to fix this. If you know how, please let me know in the comments.

Postscriptum

My computer science thesis is going to be tightly connected with this project, yet separate enough so they can be finished with different deadlines. More in the next post.