In the last 40 years, hundreds of person-years have been spent by NASA and others developing software that accurately predicts the behavior of complex thermo/fluid/structural systems. In the last 15 years, these tools have been made easier to use with graphical user interfaces and increased connectivity between codes. However, almost all of this investment has been made in point design simulation tools: given a specific design, the software predicts how the design will perform under a specific set of conditions. This approach has become so engrained that most engineers have forgotten that what?s needed is precisely the opposite: given performance requirements and goals, find the best performing design. Moreover, environmental conditions and many performance characteristics are rarely known exactly, but rather statistically: a single point design evaluation is by itself nearly meaningless. Fortunately, prior investment in infrastructure can be applied to the next generation of tools that we propose to build. Variability is the key to the success of future design tools. Parametric CAD systems are becoming the norm, but most mechanical analysis tools have not kept pace. We led the way in NASA?s standard thermal control analyzer, even introducing built-in design optimization and data correlation modules. Now it is time for the next step. We will develop a means for characterizing designs statistically, allowing a designer to understand the risks in any design along with the cost of reducing those risks. We will also develop advanced statistical design synthesis tools that can handle difficult tasks such as selection problems and complex objectives.
Potential Commercial Applications:Previously developed thermal/fluid design optimization tools are already being successfully commercialized, and concurrent developments are targeting optimized thermal/structural integration of electronic packaging. The commercial demands for high-level design optimization tools will assure the commercialization of the resulting software, including the results of the Phase I project.
