MONASH MOTORSPORT ACCELERATES DEVELOPMENT WITH ANSYS
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
ACADEMIC STUDENT TEAM MONASH MOTORSPORT ACCELERATES DEVELOPMENT WITH ANSYS The team from Monash University uses ANSYS software to design an innovative drag reduction system. By Scott Wordley, Stuart Buckingham, Damien McArthur, Marc Russouw, Luke Phersson and Matt Corallo, Monash Motorsport, Monash University, Melbourne, Australia F ormula 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 2011 Monash Motorsport car finishing the endurance event at Formula Student Germany skills in teamwork, communication and in 2012. The Monash team finished in fourth place overall. project management, along with helping them to prepare for the challenges they will face when they embark on profes- sional careers in industry. Through a close relationship with ANSYS software and present examples of Engineering students who partici- the local ANSYS channel partner, LEAP their work. pate in this program benefit greatly from Australia, the group developed a range Monash Motorsport recently final- the opportunity to develop their exper- of tutorials to help team members ized design and development of its tise in computer-aided design and engi- and other students conduct finite ele- latest racer, the M13. The team’s cars are neering (CAD and CAE) by modeling and ment analysis (FEA) and computational well known for their distinctive aerody- simulating many different components fluid dynamics (CFD) studies for com- namic packages; Monash claimed four and systems within the Formula SAE mon Formula SAE applications. Each competition wins and several top-five car. Monash Motorsport team members year over spring break, the Monash places in recent Australian, U.K. and have utilized ANSYS engineering simula- team organizes a three-day symposium German events. Aerodynamic packages tion software for more than a decade to called Design to Win, during which local in Formula SAE are becoming popular, as accomplish this work. Formula SAE teams receive training on teams learn that wings do indeed offer © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
ACADEMIC STUDENT TEAM 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 IMPROVEMENTS The new M13 racer is a clean-sheet redesign in all respects, incorporat- ing a number of significant aerody- The Monash M12 car with DRS engaged and flaps open during on-track testing 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 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. 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 ANSYS CFX results demonstrate the dramatic differences in surface pressure magnitude a more representative measure of the and vortex structures (iso-surfaces) generated by high-downforce mode (left) and car’s performance on track while taking DRS-engaged low-drag mode (right). DRS activation results in a 50 percent reduction geometry into account) of greater than in drag generated by the full car. 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 The team chose to use a symmetry The Monash Motorsport team won its fifth fluid domain and boundary setup to model (using only half the car) to maxi- consecutive Australasian FSAE Champion- ensure consistency and comparability mize mesh resolution, given RAM-based ship in 2013, with a strong performance on track and in all static events. The team will among all future simulations. Domain meshing and solution time limitations also compete with their 2013 car at Formula size and mesh sensitivity studies were when working on single local nodes. Student UK and Germany in 2014, hoping to undertaken, and benchmarking tests Testing showed that approximately improve upon third- and fourth-place finishes were conducted with different turbu- 15 million to 20 million elements for in these events in 2012. lence models. © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
ACADEMIC STUDENT TEAM 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 Colored velocity profiles — and vectors Sample ANSYS CFD-Post pressure contours at inlet to the undertray — are useful to aerodynamic design iterations over a and surface streamlines for M13 car. qualitatively compare how much mass three-month period at the start of 2013. flow is being captured by the inlet (low Yellow and red areas denote pressures A team-developed ANSYS CFD-Post state above static pressure, and blue denotes transverse velocity components shown file was used to allow fast and consis- areas below static pressure. Surface by the vectors in the plane but also high tent automated post-processing as well streamlines help denote flow separation resultant velocity magnitude shown by and re-attachment lines as well as vortex the color contours) as well as separate as output of figures (pressure contours, planes in regions of interest for automated, activity impinging on vehicle surfaces. streamlines and vortex cores), tables quantitative calculation of mass flow rates of force and coefficient results via the (for example, radiator and diffuser tunnels). 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- Semi-transparent vortex core iso-surfaces, colored for stream-wise vorticity, to indicate files as well as to better understand direction of vortex rotation. Due to use of a symmetry model and reflection of the results, span-wise pressure variations. colors are the same on either side of the car. 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- Example of chord-wise coefficient-of-pressure plots for 2012 three-element front wing. out the design phase, which facilitated Dark blue denotes the slice at the centerline of car; lighter blues are slices taken moving rapid communication and results shar- progressively outboard from center. These results demonstrate the significant contribution ing among the team. This ensured that made by the center of the wing, justifying the team’s choice to increase car nose height to all members remained updated on the maintain the wing in this region. Note that there is no centerline data for the last wing design progress, which minimized repe- element due to a two-element configuration in the center section under the nosecone. tition and duplication as well as helped © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
ACADEMIC STUDENT TEAM student engineers to discuss and incor- porate the best design features into the next round of CFD runs. INFRASTRUCTURE FOR COMPLEX SIMULATIONS Beyond 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 Steps necessary to solve CFD simulation, including software package used for each idle desktop machines in the Monash stage. All processes in the yellow box have been scripted to run on the Beowulf cluster. Engineering Computer Labs, which were By running these processes remotely, the workstation computer is freed up to do other made available for the team’s use over- tasks. The large files created for the mesh and solution are generated remotely. Users must night and on weekends. A fully auto- download the results to post-process, but transferral of large mesh files from the remote mated grid-generation outsourcing tool system is not needed. 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. 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 Dimensions of far-field domains. All dimen- point of interest and the line that radiates from the center of rotation and is perpendicular sions are non-dimensionalized by dividing to the centerline of the car. By decreasing the parameter r, both θfront and θrear increase. by total car length. 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. © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
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 Test case of empty domain (reduced to only a few elements in from other cases and should not be used for direct comparison.) height) to confirm that boundary conditions set for the domain produced expected flow WIND TUNNEL EXPERIMENTS Monash 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 Turbulent wake profile of rotating reference frame case visualized using volume render of sweeps made in freestream in the tun- turbulent kinetic energy, as viewed from right side of race car 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. Turbulent wake profile of rotating reference frame case visualized using volume render of turbulent kinetic energy, as viewed from above race car © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
Example of full car yaw sweep undertaken in wind tunnel Results of specific dissipation testing undertaken in wind tunnel 2011 Monash Motorsport car finishing the endurance event at Formula Student Germany in 2012. The Monash team finished in fourth place overall. Team engineers extensively correlate aerodynamic predictions obtained from ANSYS software On-track testing results were used to correct wind tunnel results with data from on-track testing. for the front wing. SUMMARY significant improvement in team knowl- Authors’ Note CFD has proven to be a powerful tool edge transfer (with reports of each run Monash Motorsport has published sev- for the Monash Motorsport team, par- saved on the team knowledge data- eral SAE papers on the aerodynamic ticularly since wind tunnel testing time base for future members to access and development of its past cars, and team is limited to a few days each year. The learn from) as well as ease of compar- members are happy to talk with other team can narrow down the most prom- ison between runs. The use of CFD has teams implementing aerodynamic ising design concepts without having to allowed the team to spend financial and studies. The team sincerely thanks all incur the cost of fabricating each design time resources for building and testing current and past team members for change and physically testing it in the various prototype designs on only the their hard work and dedication to this wind tunnel or on track. Furthermore, most promising few. project — as well as LEAP Australia, automation of the simulation setup in ANSYS, the Department of Mechanical ANSYS software has allowed for quicker and Aerospace Engineering at Monash, turnaround times on simulations (down and the Monash wind tunnel facility. to 12 hours from 24 hours), and stan- dardized report generation has yielded © 2014 ANSYS, INC. ANSYS ADVANTAGE Volume VIII | Issue 1 | 2014
You can also read