Ford Performance & Team Penske (2017-Present)
My tenure with Ford can be neatly divided into two roles:
- Working as an embedded aerodynamicist with Team Penske, and
- Experimental aerodynamics lead supporting factory Ford Performance race programs.
Under the second role, since early 2020 I have been heavily involved in the development of Ford’s submission for the NextGen NASCAR.
Developing Ford’s NextGen NASCAR
Each successive NASCAR body style has withdrawn from the appearance of a true “stock” vehicle. By the 4th generation, the cars bore almost no resemblance to their road-going counterparts.
However, the new 7th generation (officially called NextGen) aims to uproot previous trends by closely matching the styling of the production vehicle.
The NextGen aerodynamic platform is entirely new. The vehicle has a completely flat underbody and a common diffuser. Bodywork is no longer asymmetrical, which actually presents a challenge for aerodynamic stability.
Besides adapting to an entirely new aerodynamic platform, some other challenges have included:
-
Meeting the design studio’s stylistic restrictions and optimizing style-based features for aero performance
-
Complying with NASCAR’s very tight aerodynamic submission requirements
-
Designing for both stability and performance (two cars with similar wind tunnel readings may behave very differently on track)
To accomplish these goals, data-driven decisions are made based on simulations and experiments.
Track Testing
Arguably the most important step in the development process is track testing. Final verification must be done on track, where conditions are as close as possible to an actual race weekend.
Track testing in NASCAR is extremely limited. However, there are sanctioned opportunities for the OEMs and teams to instrument vehicles and collect data.
Track testing is difficult relative to more controlled experiments. Wind, heat, and interaction with vehicles are inevitable, and can obfuscate data if not handled correctly.
Below I discuss two aerodynamic devices that I have experience with. They are highly relevant but not limited to track testing.
Pitot Probes
Pitot static probes are a rudimentary, but highly effective tool for measuring:
- Air speed
- Air direction
While pitot probes are basic in principle of operation, their designs can be highly complex (see the Formula 1 example below). Despite potential design intricacies, all pitot tubes operate in accordance with the simple diagram below.
Many pitot designs have multiple holes and different tip shapes to accommodate varying degrees of flow angularity. These devices are also common on all modern aircraft with varying levels of complexity.
Why use a pitot tube on a car?
The diagram below outlines of how the angle of air flow perceived by a vehicle may not match it’s direction of travel.
Other factors like wind and turbulence from other vehicles affect the angle and speed at which a vehicle moves through the air, but they are beyond the scope of explanation here.
Static Pressure Taps
To measure changes in air pressure on a vehicle, tiny holes of very specific size and location are drilled into the surface of the car. Tiny tubes are connected to these holes, bundled together, and connected to a central measurement device.
While seemingly unsophisticated, ensuring data accuracy with this method requires experimental rigor and in-depth data analysis. The pressure changes measured by these tiny holes give fantastic insight into the aerodynamic performance of the car.
The hole locations themselves are often chosen to exactly match the wind tunnel, allowing a direct comparison to a more controlled experimental tool.
Wind Tunnel
Tunnel testing boils down to creating new aerodynamic concepts and determining which are best. However, the best design is not always obvious, and rigorous data analysis is required to determine which concepts will create the most performance.
I have worked through this process for multiple vehicles while at Ford, no two vehicles have the same performance goals.
Wind tunnels are ubiquitously used across the racing industry, and are the primary development tool driving aerodynamic performance. There are a huge variety of different wind tunnel types, but for race cars the two most common features are:
- A rolling belt to simulate the road speeding past under the car, and
- An actuation system to control the position and height of the vehicle.
In most wind tunnels a fan drives the air around a closed loop, thereby recycling the air. This presents a challenge in managing the turbulence of the recycled air, but is much more power efficient.
Scale Model vs. Full Scale
In some racing series, development time is split between
- Smaller scale model wind tunnels, and
- Full size ‘full scale’ wind tunnels.
Many racing series are restricted to one of the two (usually for financial or rules compliance reasons).
Scale model tunnels often use a smaller (40-60% scale), highly modular test vehicles. Therefore, scale tunnels allow fast iteration of ideas/concepts during testing. However, the physics of a scaled down car is not a perfect representation of the full size vehicle.
Full scale tunnels may be used to test a race-ready vehicle with only minor modifications.
| Pros | Cons |
|---|---|
| Full size car will deform, vibrate, and overall behave more like the on-track vehicle | Extremely expensive |
| Greater range of achievable wind speeds, can closely match aerodynamic conditions to reality | Slow iteration time compared to scale model |
| Can be used to test a race-ready car with only minor modifications | |
| Aerodynamic physics more representative of track |
| Pros | Cons |
|---|---|
| Modular model allows for fast design iteration | Flow physics inaccuracies due to smaller size and lower speeds |
| More sophisticated instrumentation (pressure, flow measurement) | Model does not fully replicate construction of the race car |
| Less energy consumption = lower cost | |
| More easily perform advanced experimental methods |
For info on even more sophisticated aerodynamic measurement tools, see my research background here.
Team Penske Aerodynamicist
My other major role with Ford Performance is supporting Team Penske while embedded in their aerodynamics department.
My major responsibilities have included:
-
Development of an end-to-end Computational Fluid Dynamics (CFD) process to automate the analysis of aerodynamic simulation data.
-
Using the above process in collaboration with other Penske aerodynamicists to optimize the performance of Penske’s NASCAR Cup series race cars.
I’m very proud to have been with Penske during their 2018 Cup Championship. During my time at Penske the team has won 24 Cup Series races, and I’m honored to work with such an upstanding and successful organization.
Prior to COVID, I spent 4/5 days per week working out of the Penske shop. Their culture of precision and attention to detail is a big part of my growth as an engineer.
Aerodynamic Optimization
While all NASCARs may look the same on TV, this is far from the truth. Keen eyes may spot subtle differences between the front-runners and slower cars.
Up close, these differences are less subtle. This was especially true in the 4th generation of NASCAR, with bodies massively twisted in the pursuit of aerodynamic performance.
To combat all the team modifications, NASCAR enforces regulations with optical technology known as Hawk-Eye. Every car is scanned to ensure the bodywork is compliant. However, teams still push the boundaries of performance while meeting the rules as written.
Each year NASCAR closes more loopholes, and performance becomes harder to find. Engineering creativity, and strong understanding of the airflow physics are now essential.
Often, the most innovative ideas are demarcated by their impact on the rulebook (i.e the best ideas are quickly banned once discovered).
Modern simulation tools allow us to develop a deep understanding of the car’s aerodynamics. Without simulation, it would be exponentially more difficult to squeeze out the last bit of performance.
Simulating Air: CFD
Computational fluid dynamics (CFD) uses brute force to solve the equations that govern fluid flow. Turbulence makes simulating fluids very difficult, and even with modern computers, models and approximations are necessary.
CFD allows engineers to visualize air flow without massively complex measurements. Even if the models are not completely correct, the insight gained from CFD sparks new design directions and optimizations that are impossible with experiments alone.
Many days of my career have been spent digging through CFD data. CFD is the best tool for generating creativity and scheming new ideas that generate downforce and reduce drag.
I’m also always looking to develop more efficient work flows that condense datasets for easy visualization (important for idea generation).
With simulations you can even directly look at the difference between two aerodynamic flow features, and theorize better ways to achieve goals based on the structure of the air.
While methods are less often discussed, they are extremely important to standardize work flows and analysis. Every simulation yields a quantity of data that makes it impossible to look at every facet of the results. Therefore, every organization or team will develop their own philosophies for standardization of analysis.
These methods (and their outputs) usually involve proprietary scripts and software. Organizations protect their methods and analysis format tightly, as they have historically been differentiating factors. Former F1 technical director William Toet give a rare candid look at some basic CFD methods here.
The choice of simulation turbulence model is actually less important than academia might emphasize. Rather, a strong understanding of the shortcomings of any chosen model will be more fruitful in determining how it fits into the design process.
Audi Sport (2015)
I was fortunate to spend six months working at Audi Sport as a Praktikant (German internship/placement program), helping the team develop the 2016 R18 LMP1 race car.
I primarily assisted with development of their CFD methods and workflow, but due to my strong performance, I was allowed to contribute to the aerodynamic optimization process at the later portion of my placement.
The low drag variant of the 2016 R18 LMP1 is pictured below. The R18 utilized a huge array of advanced technology, even by today’s standards.
CFD Methods and Work Flows
In order to maximize throughput, race teams and other industrial CFD users will automate setup and analysis processes with custom scripts, apps, and dashboards.
While at Audi, I developed scripts in VBA, Python, Matlab and Java that improved workflow efficiency of the engineering team.
Some of the tools I worked on were basic, but significantly reduced manual input. Others were more complex, and one even attempted to use graph theory to help detect surface tessellation errors.
Aerodynamic Optimization
The aerodynamic optimization of the R18 led to some highly complex body surfaces. While my involvement was small, I got to learn a vast amount about practical aerodynamics and how to manipulate air flows.
This section exemplifies some of the bodywork complexity arising from collective optimization efforts.
The 2015 and 2016 R18 were unique in that they allowed a high flow rate of air to pass through the car (rather than over).
Particularly, the 2016 had a raised skinny nose that produced a large open area above the front wing, maximizing air flow through the car. This potentially allows for more rear downforce to be produced, but could also add significant parasitic drag. Hence the complexity in some of the components normally hidden from sight.
Turning vanes take some of the turbulent air from the front wing, and push it out and away from the car. They also produce vortices that generate load (downforce) on the leading edge of the floor. More vortices meant more front downforce, but at the cost of sending turbulent air to the diffuser.
LMP1 cars at the time had complex hybrid systems, and were turbocharged on the Audi and Porsche. The engines had to reliably operate under race conditions for 24 hours. Therefore, engine cooling demands resulted in some unique design choices.
Moreover, the Mulsanne straight at Le Mans forces cars from 200+ mph into low speed cornering conditions. These demanding braking scenarios led to complex brake ducts similar to Formula 1.
Due to the VW diesel scandal and the high cost of LMP1, Audi raced its final season in 2016.
However, the 2017 car was rumored nearly complete. A scale wind tunnel model was photographed by former driver Lucas di Grassi. Plans for further increases in complexity are apparent.
Ontario Power Generation (2013-2014)
I began my career at Ontario Power Generation, working as an intern from May 2013 to June 2014. I was positioned at the Pickering Nuclear Generating Station.
My role was as a systems engineer for the plant’s automated security, working under the senior engineer responsible for the system.
The intrusion detection system was particularly complex, and offloaded large datasets for diagnostic purposes. I primarily developed code to analyze these datasets, and concatenate relevant metrics into regulatory reports. Security standards are set out by the Nuclear Regulatory Commission (NRC), and they require such reporting to ensure system reliability.
As part of the regulatory report, thousands of signals needed to be dissected, cleaned, and summarized. I completed the development of a VBA based tool that automated the regulatory reporting that is still in use today. I also wrote a user manual, and initiated procedures to permit the code to be used as an official company tool.
As a side benefit, I got to learn a lot about Nuclear Power and conventional steam cycles.
Driven by my curiosity, I used all the available plant drawings to create my own hand-drawn schematic of the major components in the plant’s Rankine cycle. Myself and several colleagues then used the schematic as a map to visit each component in-person.
