Overview: Modern aerospace systems are increasingly complex and multidisciplinary,
and are required to achieve dramatic improvements in performance, cost effectiveness,
and reliability. The field of multidisciplinary optimization (MDO) provides
many tools for design optimization within the framework of decomposed and
re-combined systems. There is also a suite of probabilistic methods that may
be used to assess the reliability of general systems. Synthesizing these tools
in an integrated design process, however, exacerbates the already difficult
challenges regarding computational complexity and poor convergence. This research
therefore seeks to formulate an efficient probabilistic multidisciplinary
optimization strategy, which combines MDO techniques, reliability analysis,
and reliability-based optimization methods. The optimization and reliability
analysis iterations are decoupled (compared to earlier nested methods) to
achieve computational efficiency. In addition, a systematic methodology is
sought that uses model error sensitivity as a basis for selecting the fidelity
level of disciplinary models. The methodology is to be demonstrated for the
integrated hypersonic aeromechanics tool (IHAT), which has been developed
to aid in designing hypersonic missile systems for the U.S. Navy. The research
is conducted in close collaboration with researchers at NASA
Langley Research Center.
The multidisciplinary probabilistic optimization methodology is extended to integrate system-level and component-level design requirements. Large aerospace systems are too complex to study on a single level, so a top-down design hierarchy is needed. In this approach the system is first designed with minimal detail using with low-fidelity analyses; during subsequent design stages, components are designed with higher fidelity analyses. However, if the design is to be most efficient, the system-level design must be updated considering additional detail provided from component design. The coupling of the two design levels is modeled through an efficient multidisciplinary optimization formulation.
Example Applications: The proposed methods are being demonstrated
on multidisciplinary analysis models for aerospace design applications.
 |
 |
|
Reusable
Launch Vehicle Concept
|
Air-launched
low volume ramjet
|
Potential Applications: The incorporation of probabilistic information in multidisciplinary design is a valuable tool for numerous applications across all engineering disciplines. In fact, as systems become more complex and continue to incorporate new technologies, limitations on funding and time may necessitate this modeling and simulation-based approach.