New computing tools, international collaboration spell design progress

By Douglas L. Allaire and Karen E. Willcox|December 2018

The Multidisciplinary Design Optimization Technical Committee provides a forum for those active in development, application and teaching of a formal design methodology based on the integration of disciplinary analyses and sensitivity analyses, optimization and artificial intelligence.

OpenMDAO is an open-source high-performance computing platform for systems analysis and multidisciplinary optimization. In June, the optimal control library Dymos was released and used to optimize electric aircraft trajectories with acoustic and thermal constraints. Dymos is built on OpenMDAO Version 2, which NASA released late in 2017. Also in June, a collaboration between NASA and the University of Michigan used OpenMDAO to develop the first fully coupled aeropropulsive MDO for a boundary layer ingestion propulsion system.

In January, Mississippi State University released the Multidisciplinary-design Adaptation and Sensitivity Toolkit multiphysics, an MDO toolkit that supports coupled conduction, nonlinear structural, fluid flow, and flutter analyses, and sensitivity with respect to sizing and level set topology parameters. In July, University of California, San Diego published open-source level set topology optimization software. This new software routinely runs 10 million element models by adapting and tailoring the level set method, making design for additive manufacturing immediately accessible.

In January, NASA’s Goddard Space Flight Center in Maryland developed an MDO framework for design under uncertainty and model validation of the James Webb Space Telescope. The European-funded H2020 project launched the AGILE Academy, AGILE being short for Aircraft Third Generation MDO for Innovative Collaboration of Heterogeneous teams of Experts.The academy hosts more than 40 students from 15 institutions across the world from February to July. In January, the University of Illinois developed new system architecture design methods that use machine learning with data from enumeration and design optimization, with application to aircraft cooling systems and power electronics. In March, the University of Michigan performed the high-fidelity aerostructural design optimization of a tow-steered composite high aspect ratio wing. A third scale model of the optimized wing box was built by Aurora Flight Sciences using an automatic fiber placement machine and will undergo ground vibration testing at NASA. In June, Embraer and Instituto Tecnológico de Aeronáutica in Brazil developed new methods for optimization under uncertainty and demonstrated how upfront reduction of uncertainties improves design competitiveness without sacrificing reliability.

M3 is a U.S. Air Force Multidisciplinary University Research Initiative developing new methods to manage multiple information sources in MDO. In July, the M3 team released BOCS, a toolbox for Bayesian Optimization of Combinatorial Structures, and in November the team released a multifidelity toolbox for evaluating stability boundaries, CLoVER, short for Contour Location Via Entropy Reduction.

Boeing completed the Air Force Research Laboratory’s Optimized Integrated Multidisciplinary Systems program in October. OPTIMUS showed the importance of including power and thermal subsystems during conceptual-level design, not only for understanding the internal subsystems themselves, but also for their pronounced effects on the outer mold line and shape of the air vehicle. This program also pushed the state-of-the-art in distributed collaborative design that was demonstrated between Boeing and GE.

Photo: Aurora Flight Sciences’ tow-steered wing skin is fabricated in Columbus, Mississippi. Credit: Aurora Flight Sciences