Common Misconceptions in Aerodynamics: Part 3

The first part of this series dealt with the erroneous use of "templates" to guide aerodynamic modifications, and the second part with the belief that testing is unnecessary or impossible to do at home or both. Now, we'll take a look at the prevailing belief that aerodynamics is an easy subject.

Aerodynamics are Extremely Complicated

The claim: The principles of fluid dynamics that govern airflows are simple to understand; therefore, aerodynamics as a whole are actually simple. Anyone who says or thinks aerodynamics are complicated is just muddying the waters.

There's so much wrong in this forum post, I'm not even sure where to start.

The reality: Hand-in-hand with the idea that airflows can be predicted and low drag is as simple as following a template, the belief that aerodynamics are easy to understand (a prerequisite for their ease of prediction) is just as erroneous. The aerodynamic force exerted on a body like a car arises from the pressures acting on it. Pressure times area gives a force; integrating all the tiny little forces created by the airflow everywhere on the body sums to a resultant force acting on an imaginary point called the “center of aerodynamic pressure,” and this force can be divided into components along the three dimensions: drag/thrust (x), lift/downforce (z), and sideforce (y). The force arises from pressures acting on every surface of the body, ie its “wetted area” or everything exposed to the flow.  As AJ Scibor-Rylski wrote in Road Vehicle Aerodynamics (1975)—one of the first books dealing with vehicle aerodynamics specifically— “The aerodynamic force is the net result of all the variably distributed pressures which the airstream exerts on the car surface.”

As I’ve written elsewhere, drag does not arise from air molecules “hitting” the front of a car. This is an intuitive explanation of fluid flows that is incorrect (if you previously thought this, congratulations! You’re in good company—so did Isaac Newton). In fact, it does not arise from micro-level material properties at all, but instead from continuum-level behavior of fluids, as Doug McLean argues in Understanding Aerodynamics: Arguing from the Real Physics. These continuum-level behaviors are described by the Navier-Stokes equations, a set of simultaneous partial differential equations relating pressure, velocity, temperature, density, and other fluid properties as they change over time and in three-dimensional space. If you aren’t familiar with systems of simultaneous partial differential equations, they are impossible for humans to solve without computers crunching away at numbers for a long time, generating solutions by iterative brute force (this is essentially what computational fluid dynamics simulations are). Even then, computer programmers have to make simplifying assumptions about some variables in order to solve for a likeness of real air flows, which are in reality even more complex. This is anything but simple.

On a higher level, even predicting the change in airflow by altering a body geometry parameter is next to impossible without long experience with similar bodies and similar changes, and even then longtime aerodynamicists can be surprised by the behavior of a particular car. For example, JP Howell reported, in Sustainable Vehicle Technologies (2012), testing an Audi A2 in two different wind tunnels and finding in both that lowering the ride height did not reduce drag—something that surprised him since it does on most cars and is often repeated as a “rule of thumb” online. Because of the particular shape of the A2, its aerodynamics didn't respond in the same way as a lot of other cars--something that is only discoverable through testing.

If you’re ever feeling masochistic, have a read through the last few chapters of Scibor-Rylski’s Road Vehicle Aerodynamics. Scibor-Rylski was one of the first researchers to attempt mathematical modeling of the aerodynamic forces acting on a car traveling over a curved section of road in a constant wind. Make sure you review your calculus and engineering dynamics first, have a pencil and paper handy, be familiar with terms like “slip angle” and “steering angle” and the difference between them, understand the difference between forces and moments and have a solid theoretical grasp on the three dimensions of both linear forces and moment axes, and be familiar with the author’s specific variable scheme to represent all these. Sound easy? It isn’t. And this is just a simplified mathematical representation of a single case.

My attempt to follow along with Scibor-Rylski's model. Note the unrealistic assumptions, such as CG and CP height being the same.

As the adage says, if aerodynamics were easy, everyone would be doing it. Unfortunately, what we see today in online communities are lots of people who think it is easy because they’ve been misled. Don’t get sucked in by them.