ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

ACADEMIC STUDENT TEAM

MONASH MOTORSPORT ACCELERATES DEVELOPMENT WITH ANSYSThe team from Monash University uses ANSYS software to design an innovative drag reduction system.

Formula SAE competition chal-lenges university students each year to design, build, market and race a small, open-wheeled, Formula-style car against other such insti-

tutions from around the globe. Monash Motorsport, from Monash University in Melbourne, Australia, is one of these teams. The Monash Motorsport team com-prises approximately 70 under-graduate students, primarily from the Department of Mechanical and Aerospace Engineering but also from disciplines such as science, business, marketing and even law. The demand-ing nature of the competition gives stu-dents the chance to develop important skills in teamwork, communication and project management, along with helping them to prepare for the challenges they will face when they embark on profes-sional careers in industry. Engineering students who partici-pate in this program benefit greatly from the opportunity to develop their exper-tise in computer-aided design and engi-neering (CAD and CAE) by modeling and simulating many different components and systems within the Formula SAE car. Monash Motorsport team members have utilized ANSYS engineering simula-tion software for more than a decade to accomplish this work.

Through a close relationship with the local ANSYS channel partner, LEAP Australia, the group developed a range of tutorials to help team members and other students conduct finite ele-ment analysis (FEA) and computational fluid dynamics (CFD) studies for com-mon Formula SAE applications. Each year over spring break, the Monash team organizes a three-day symposium called Design to Win, during which local Formula SAE teams receive training on

By Scott Wordley, Stuart Buckingham, Damien McArthur, Marc Russouw, Luke Phersson and Matt Corallo, Monash Motorsport, Monash University, Melbourne, Australia

2011 Monash Motorsport car finishing the endurance event at Formula Student Germany in 2012. The Monash team finished in fourth place overall.

ANSYS software and present examples of their work. Monash Motorsport recently final-ized design and development of its latest racer, the M13. The team’s cars are well known for their distinctive aerody-namic packages; Monash claimed four competition wins and several top-five places in recent Australian, U.K. and German events. Aerodynamic packages in Formula SAE are becoming popular, as teams learn that wings do indeed offer

ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

ACADEMIC STUDENT TEAM

The Monash Motorsport team won its fifth consecutive Australasian FSAE Champion-ship in 2013, with a strong performance on track and in all static events. The team will also compete with their 2013 car at Formula Student UK and Germany in 2014, hoping to improve upon third- and fourth-place finishes in these events in 2012.

benefits! Successful implementation of a well-designed aero package can result in significant improvements in a car’s per-formance on tight, twisting Formula SAE tracks, even though the average track speed is low — at around 60 kilometers per hour.

AERODYNAMIC IMPROVEMENTSThe new M13 racer is a clean-sheet redesign in all respects, incorporat-ing a number of significant aerody-namic improvements and novel design features made possible by extensively using ANSYS Mechanical, ANSYS CFX and ANSYS Fluent. The car is one of only a few Formula SAE vehicles world-wide to utilize a drag-reduction system (DRS), used in current Formula One rac-ing. This innovation enables the angle of the flaps in the multi-element front and rear wings to be dynamically adjusted via pneumatic cylinders and linkages. As a result, the car has two distinct aero-dynamic modes: high downforce and low drag. The DRS is activated when the driver presses a button on the steering wheel, so the low-drag setting can be used when the car is driving in a straight line and significant downforce is not required. By using a button to engage DRS, the driver can revert to the high-downforce mode and maximize the car’s downforce (and drag) before applying the brakes at the end of a straight — which an automated system cannot do reliably without GPS-enabled track mapping. Having access to a drag-reduction system has allowed this year’s Monash team to significantly increase its down-force target for the M13 car to a CL.A (“A” denotes frontal area — when com-bined with lift coefficient, this provides a more representative measure of the car’s performance on track while taking geometry into account) of greater than 6, given that drag is no longer a signif-icant limitation on straight-line perfor-mance. Before starting design, the team invested time in developing a standard fluid domain and boundary setup to ensure consistency and comparability among all future simulations. Domain size and mesh sensitivity studies were undertaken, and benchmarking tests were conducted with different turbu-lence models.

The team chose to use a symmetry model (using only half the car) to maxi-mize mesh resolution, given RAM-based meshing and solution time limitations when working on single local nodes. Testing showed that approximately 15 million to 20 million elements for

ANSYS CFX velocity contours and vectors show how the different flap rotation angles for DRS were tested and optimized, using a 3-D CFD model of the entire car.

The Monash M12 car with DRS engaged and flaps open during on-track testing

ANSYS CFX results demonstrate the dramatic differences in surface pressure magnitude and vortex structures (iso-surfaces) generated by high-downforce mode (left) and DRS-engaged low-drag mode (right). DRS activation results in a 50 percent reduction in drag generated by the full car.

ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

the symmetry model provided the best compromise between mesh resolution and solver time, based on Monash’s current computational resources. A 400-iteration run using the k-omega SST turbulence model generally solved in less than six hours, which was consid-ered an acceptable turnaround time. The eight-person aero team conducted and documented almost 200 unique aerodynamic design iterations over a three-month period at the start of 2013. A team-developed ANSYS CFD-Post state file was used to allow fast and consis-tent automated post-processing as well as output of figures (pressure contours, streamlines and vortex cores), tables of force and coefficient results via the report function. The team utilized cus-tom pressure color scales to clearly dif-ferentiate positive pressures (yellow to red) from suction pressures (blue). A custom red/blue scale was applied to stream-wise vorticity and used to color vortex core iso-surfaces, neatly high-lighting the direction of vortex rotation. Chord-wise plots of coefficient of pres-sure for the front and rear wings at a range of span-wise locations were rou-tinely generated to fine-tune wing pro-files as well as to better understand span-wise pressure variations. Keyframe animations were used extensively to generate longitudinal total pressure sweep videos and vortex core videos, providing insight into the complex vortex and wake interactions that dominate the vehicle’s near field. Juggling these vortex and wake interac-tions proved crucial in maximizing the downforce produced by the front wing and underbody diffuser, as well as in balancing the front and rear downforce distribution for the whole car. These full-car interactions drove the team’s final choice of rear-wing height and pro-vided confidence with respect to the cooling flows entering the radiator and turbo intercooler. The full results from each run, along with associated CAD models and ANSYS Workbench archives, were updated to a private team Wiki in real time through-out the design phase, which facilitated rapid communication and results shar-ing among the team. This ensured that all members remained updated on the design progress, which minimized repe-tition and duplication as well as helped

ACADEMIC STUDENT TEAM

Sample ANSYS CFD-Post pressure contours and surface streamlines for M13 car. Yellow and red areas denote pressures above static pressure, and blue denotes areas below static pressure. Surface streamlines help denote flow separation and re-attachment lines as well as vortex activity impinging on vehicle surfaces.

Colored velocity profiles — and vectors at inlet to the undertray — are useful to qualitatively compare how much mass flow is being captured by the inlet (low transverse velocity components shown by the vectors in the plane but also high resultant velocity magnitude shown by the color contours) as well as separate planes in regions of interest for automated, quantitative calculation of mass flow rates (for example, radiator and diffuser tunnels).

Semi-transparent vortex core iso-surfaces, colored for stream-wise vorticity, to indicate direction of vortex rotation. Due to use of a symmetry model and reflection of the results, colors are the same on either side of the car.

Example of chord-wise coefficient-of-pressure plots for 2012 three-element front wing. Dark blue denotes the slice at the centerline of car; lighter blues are slices taken moving progressively outboard from center. These results demonstrate the significant contribution made by the center of the wing, justifying the team’s choice to increase car nose height to maintain the wing in this region. Note that there is no centerline data for the last wing element due to a two-element configuration in the center section under the nosecone.

ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

student engineers to discuss and incor-porate the best design features into the next round of CFD runs.

INFRASTRUCTURE FOR COMPLEX SIMULATIONSBeyond conducting CFD analysis, the Monash team developed a methodology and hardware infrastructure required to conduct large and complex simulations (incorporating up to 200 million ele-ments) combined with a rotating refer-ence frame. A rotating reference frame is needed for modeling aerodynamic effects when the car is turning a corner, since the interactions cannot be accu-rately estimated nor understood using traditional fixed-flow yaw angles applied to the entire car (as in a wind tunnel). The team developed a 100-node local Beowulf-style cluster by utilizing idle desktop machines in the Monash Engineering Computer Labs, which were made available for the team’s use over-night and on weekends. A fully auto-mated grid-generation outsourcing tool was scripted to allow geometry clean-up, surface and volume meshing, and solving to be completed remotely on the cluster, thereby avoiding RAM lim-itations and slow transfer times for the large meshes, which otherwise would be generated locally.

ACADEMIC STUDENT TEAM

Steps necessary to solve CFD simulation, including software package used for each stage. All processes in the yellow box have been scripted to run on the Beowulf cluster. By running these processes remotely, the workstation computer is freed up to do other tasks. The large files created for the mesh and solution are generated remotely. Users must download the results to post-process, but transferral of large mesh files from the remote system is not needed.

The incident angle, θ, is the angle that the freestream air makes with the car centerline at the point of impact. The freestream vector is tangential to the center of rotation and, therefore, perpendicular to any line that radiates from the center of rotation. The angle is identical to that formed between the line radiating from the center of rotation to the point of interest and the line that radiates from the center of rotation and is perpendicular to the centerline of the car. By decreasing the parameter r, both θfront and θrear increase. Increasing the parameter ψ moves the center of rotation point rearward. This has the effect of reducing the rear incidence angle, θrear, but increasing the front incidence angle, θfront. Due to the large cost in time of setting up, solving and post-processing a rotating reference frame simulation on the cluster, only one case was considered. The 200 runs conducted by the team were completed for a straight line case, and automation allowed runs to be turned around in 12 hours. Using this method, the team could cycle through many different iterations within a very narrow design window of approximately three months.

Dimensions of far-field domains. All dimen-sions are non-dimensionalized by dividing by total car length.

ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

Turbulent wake profile of rotating reference frame case visualized using volume render of turbulent kinetic energy, as viewed from above race car

WIND TUNNEL EXPERIMENTSMonash Motorsport is fortunate to have access to a full-scale automotive wind tunnel on campus; this has allowed the team’s engineers to extensively correlate aerodynamic predictions obtained from ANSYS software with data from on-track testing. Typically, the team starts by cor-relating wind tunnel and CFD results for performance of wings in isolation, via wing angle-of-attack and yaw angle sweeps made in freestream in the tun-nel. Then the car is added, which allows a detailed study of the rear wing height along with endplate size, shape and detail features. Cooling performance is measured using a specific-dissipation test rig within the tunnel, which circu-lates heated coolant through the radiator at a measured flow rate. Thermocouples in the coolant lines allow the team to calculate heat dissipation as a function of the temperature differential between ambient air flow and coolant.

Test case of empty domain (reduced to only a few elements in height) to confirm that boundary conditions set for the domain produced expected flow

Turbulent wake profile of rotating reference frame case visualized using volume render to turbulent kinetic energy (TKE) measured in J/kg. The edges of the ground plane indicate the bounds of the fluid domain. This type of representation serves to illustrate the asymmetry of the wake behind the car in the rotating reference frame. (Rendering palette used in this visualization is different from other cases and should not be used for direct comparison.)

Turbulent wake profile of rotating reference frame case visualized using volume render of turbulent kinetic energy, as viewed from right side of race car

ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014 © 2014 ANSYS, INC.

Results of specific dissipation testing undertaken in wind tunnel

Example of full car yaw sweep undertaken in wind tunnel

SUMMARYCFD has proven to be a powerful tool for the Monash Motorsport team, par-ticularly since wind tunnel testing time is limited to a few days each year. The team can narrow down the most prom-ising design concepts without having to incur the cost of fabricating each design change and physically testing it in the wind tunnel or on track. Furthermore, automation of the simulation setup in ANSYS software has allowed for quicker turnaround times on simulations (down to 12 hours from 24 hours), and stan-dardized report generation has yielded

significant improvement in team knowl-edge transfer (with reports of each run saved on the team knowledge data-base for future members to access and learn from) as well as ease of compar-ison between runs. The use of CFD has allowed the team to spend financial and time resources for building and testing various prototype designs on only the most promising few.

2011 Monash Motorsport car finishing the endurance event at Formula Student Germany in 2012. The Monash team finished in fourth place overall.

Authors’ NoteMonash Motorsport has published sev-eral SAE papers on the aerodynamic development of its past cars, and team members are happy to talk with other teams implementing aerodynamic studies. The team sincerely thanks all current and past team members for their hard work and dedication to this project — as well as LEAP Australia, ANSYS, the Department of Mechanical and Aerospace Engineering at Monash, and the Monash wind tunnel facility.

On-track testing results were used to correct wind tunnel results for the front wing.

Team engineers extensively correlate aerodynamic predictions obtained from ANSYS software with data from on-track testing.