Student: Andrew Thelen, Graduate Student in Aerospace Engineering, Iowa State University
Faculty Advisor: Leifur Leifsson
Efficient design optimization using variable-fidelity optimization
Computational modeling plays a vital role in the aircraft design process. In particular, tools such as computational fluid dynamics (CFD) and finite element analysis (FEA)
are often used to estimate an aircraft’s aerodynamic performance and structural response. When compared to wind tunnel or flight testing, these methods can provide comparable information, often with minimal error, at a fraction of the time and monetary cost. In addition, simulations are often automated by parameterizing the aircraft geometry, enabling the application of numerical optimization methods. This allows the design engineer to simply initiate the optimization process, and later come back to a cheaper, more efficient, or higher performing aircraft design.
At the same time, computational models can be formulated in a variety of ways, and there is usually a trade-off between accuracy and time required per simulation. As a result, some aeronautic systems are too costly to optimize. By using a combination of faster, less accurate (low-fidelity) models and more time-consuming models of higher accuracy (high-fidelity), a system can potentially be optimized in much less time. This type of method, known as variable-fidelity optimization, is my research focus at Iowa State. In the future, I hope to implement new and existing variable-fidelity optimization methods and compare them to state-of-the-art methods.