University of Toronto's Institute of Aerospace Studies Refines Aircraft Design

Graduate Students Use SGI Computing Power to Design More Fuel-Efficient Airfoils for Aircraft of the Future — Graduate students at the Institute of Aerospace Studies (IAS) at the University of Toronto are using a 32-processor SGI Altix 3700 compute server to create wing designs with the potential of reducing drag and increasing fuel efficiency on commercial aircraft. Using computational fluid dynamics (CFD) coupled with a detailed structural model, they are finding ways to improve the interaction between the airflow over the wing and the wing itself by optimizing the shape of the wing. The Institute uses technology from Silicon Graphics (OTC: SGID) and multidisciplinary design optimization (MDO) techniques to create wing designs that satisfy multiple requirements such as lift, strength, and stability. "We need the power of the Altix system to perform high fidelity design optimizations," said Professor Joaquim Martins. "You don't need much computational power to do basic aircraft design, but if you want to be more precise and model a more refined shape, you may need to use hundreds of variables to represent variation in the shape of the airfoil at many stations along the wing, the details of the internal structure, and other information. You would then perform a high fidelity analysis of the air flow using CFD." Airfoil design studies have historically been constrained by widely accepted standards of wing thickness based on conventional structures. For that reason, design optimization has not included the possibility of altering aircraft structure. Martins has taken design optimization a step further, using MDO modeling techniques that enable his students to vary both wing structure and airfoil characteristics to optimize overall design qualities such as fuel economy. "We had worked with SGI systems previously, and we wanted the Linux environment," said Martins. "The shared memory architecture of the Altix system is useful to us because some of the tools we use require a lot of RAM. It's also convenient to be able to submit jobs to a single-address architecture as opposed to logging into multiple processors. And in my experience SGI requires a lot less maintenance than other alternatives — it's very rare to have a hardware or software failure. We've had other SGI shared-memory servers, and we've also had clusters, but the SGI Altix system is easier to manage and maintain." The SGI Altix server is currently accessed by 16 IAS faculty members and up to 80 graduate students who develop code for aerospace and aeronautical engineering design projects. Among the projects Professor Martins and his team have used the SGI Altix for was the optimization of a supersonic business jet configuration similar to the one currently being developed by Aerion Corporation. This is an aircraft that features laminar flow wings and can cruise up to 1.6 Mach. IAS installed the SGI Altix 3700 server powered by 32 Intel Itanium 2 processors in 2004. More information about the Institute of Aerospace Studies at the University of Toronto can be found on the web at its Web site.