Momentum - October 2020 - 4

STUDENT

GENERATION

FEATURE

DITCHINGTHE BELLS
AND WHISTLES

The Northwestern Formula Racing car rounds a corner on its first and only lap of the
Endurance Event at Formula SAE 2019. Problems with shifting and a failed ground
wire were the reliability issues holding the team back from competitive success.

How the Northwestern University Formula SAE team improved their car's
aerodynamics package without expensive conventional technology.
"VALIDATE!!!!!,"with the bold font, capital letters, and
exclamation marks all included, was the main feedback design
judges gave the Northwestern Formula Racing (NFR)
aerodynamics subteam at the Formula SAE Michigan
competition in 2019. At least, that's how it shows up in my
notes. The judge who explained our score to us was just about
begging us to validate the CFD we did on the aero package by
gathering empirical data on the real car, since our operating
principle has always been to scratch our heads and say "good
enough." He then walked us through how to prepare for wind
tunnel testing, how to get the most out of it, and how to
correlate the data we gathered to our simulated results.
Validation has been an issue for us since the team first
added an aerodynamics package for NFR2016. Previous teams'
idea of aero design was to pick out airfoils with good-looking
lift/drag graphs, read about the theoretically best wing
configurations, then run different cases in 2D CFD until the
velocity graphs looked pretty. They had the equivalent of a
kid's license to the CFD software they were using, limiting them
to seven iterations for any 3D scenario they wanted to test.
Since values for downforce and drag start to converge in the
thousands of iterations, those seven iterations they had can be
just about guaranteed to be so far from real values as to be
completely useless. Their downforce came from the weight of

4 October 2020

the resin-rich carbon fiber. Validation wasn't even a concern for
them since they didn't even have reliable 3D CFD
measurements to validate against.
Our first try at actually validating our CFD was last year, for
NFR19, when the team commissioned a Capstone team to
redesign our woefully inadequate aerodynamics package. They
found data gathered using a wind tunnel for lift/drag
characteristics of a NACA 0012 airfoil at different angles of
attack, then replicated those test cases in 3D CFD-they finally
had a full license to the software so they weren't limited to just
seven iterations- and found the results were close enough that
they could consider the CFD validated for all scenarios. They
designed their aero package with a simplified model of the car
in 3D, balancing the center of pressure (CP) location by detuning the rear wing to bring the CP forward. In their
simulation template, their virtual wind tunnel was roughly three
times in height, three times in width, and six times in length
larger than the car, compared to the respective six, ten, and
thirty times the car's measurements recommended as
minimum sizes for the wind tunnel to minimize boundary
interference with flow. They also did not consider boundary
conditions interfering with flow in their validation, so they
didn't scale the wind tunnel according to the size of the body
within it. Any larger tunnel could not be meshed and run

MOMENTUM
Momentum - October 2020

Table of Contents for the Digital Edition of Momentum - October 2020
