Optimization propelling the M-Fly SAE Aero Design Team

M-Fly is the University of Michigan's SAE Aero Design team, a student-led project with a mission: design, build, and make an R/C aircraft fly.

The SAE Aero Design Competition was created to connect engineering students with a real-life engineering experiences and prepare them for their professional paths. Annually, M-Fly participates in the Regular Class of the SAE Aero Design Competition. For the season 2016/2017, the competition objective is to maximize the amount of "passengers" without leaving empty seats, a realistic challenge faced by commercial airliners. In fact the airplane shall lift the heaviest payload possible, takeoff within 200 feet, complete one loop and land within 400ft. In addition to that, starting from 2016, M-Fly also competes in the Advanced Class of the SAE Aero Design Competition. The Advanced Class includes the design of the internal combustion power, static and dynamic payloads - that must be dropped on a target during the flight - and the use of sensors and other electronic systems.

Due to the long and freezing winter in Michigan, the M-Fly team just has few weeks left to design and test the plane. M-Fly has partnered with ESTECO Academy and will benefit from free training and access to modeFRONTIER optimization platform to improve their aircraft design and validate analysis results, hopefully faster.

"At M-Fly, our goal is to teach aerospace engineering, specifically aircraft design through competing at the SAE Aero Design competition. We balance winning and teaching, so we try to involve as many interested University of Michigan students in our project, while still designing the best aircraft for the competition. However, our design cycle is brutal as we have two major factors against us: our school year and the weather. Our school year starts in September which means that we, after classes, need to recruit new members and train them; for this reason we are forced to start the design and testing phase around early October. That means if we want to finish the testing phase before we head to competition, we need to get to the final design by Thanksgiving (mid,late-November) and finish the construction in January: this is a very tight schedule. In Michigan, from December through March, the temperature highs are hovering at freezing temperatures and opportunities for prime weather conditions to flight test are minimal. If we get lucky, we can perform a flight test or two before we head off to competition which is in the much nicer Southern United States (the competition rotates between Florida, Georgia, and Texas). The more time we have with a full aircraft built, the much more increased chances of us getting more test flights in. Therefore, first, we only have a month and a half to design the plane and, second, we need to analyze and test as much as we can the individual components and systems of our aircraft so we make sure that they are highly reliable to offset the lack of flight test opportunities.

That is where modeFRONTIER comes into play. modeFRONTIER allows us to explore a much larger design space in significantly less time than we could do by ourselves. Just the Design of Experiments (DOE) runs give us more data that we have ever gotten in our past design cycles in terms of different configurations.

We are currently using modeFRONTIER to do two things: iterate through many different configurations to optimize and do multi-disciplinary analysis since it interfaces so well with other analysis and CAD software we have here at Michigan such as ANSYS, StarCCM+, and SolidWorks. Instead of a standard design, analyze, build, test, go back to first step and repeat - design cycle, we can multiply the iterations for each step: design x 10000 -> analyze x 10000 -> downselect design -> build -> test and repeat the last 3 steps, with the first 3 steps taking only a couple hours or a single night if needed. modeFRONTIER also has a superb post processing capability that allows us to analyze our results in many different ways to make sure we are choosing the right design, as well as provide insights into our design problem. We solve both our problems: we earned more time to design and we are now able to validate our analysis results.

Currently, we are using modeFRONTIER for wing geometry optimization through AVL and tail aerostructural optimization. We are also working on getting high fidelity analysis/optimization (CFD and FEA). We are confident that we can get it working because modeFRONTIER's GUI is very easy to work with and interfaces very well with other major reputable analysis software."

Beldon Lin 

(member of the M-Fly SAE Aero Design Team
and Advanced Class Chief Engineer at the University of Michigan)