Software Integrates Disciplines for Aircraft Design
In the aircraft design industry, the area of geometry and geometry handling is considered the least glamorous part of the design and analysis process. But the “ugly stepchild” of the design process also usually contributes to 75 percent of the elapsed time of the project.
So why is this process so unloved? According to College of Engineering and Computer Science Associate Professor John Dannenhoffer, the answer comes down to a lack of recognition for its complexity and required rigor. “We have to take a CAD design and get rid of all of the ‘sins’ so we can create something that’s beautiful for the computational fluid dynamics (CFD) people. Its ugly, but it has to be done to get good solutions.”
With funding from NASA over the years, Professor Dannenhoffer has collaborated with a researcher from Massachusetts Institute of Technology, Bob Haimes, to develop the tools and techniques to tackle this problem and to make the transition process between disciplines more efficient and accurate.
Dannenhoffer Receives Best Paper Award
“In airplane design, different disciplines have different priorities. Structures engineers want to keep the holes in a design because they care about stress concentrations, but CDF engineers want to plug the holes because it provides information that clutters their analysis. Setting things up so that there is communication between the disciplines takes a considerable amount of time.”
Also, since CFD and structural analysis use different approximations, there is concern that when the data is transferred small errors will accumulate with the chance of contaminating the whole process.
A paper addressing this issue by Dannenhoffer and Haimes, “Conservative Fitting for Multi-Disciplinary Analysis” was awarded the Shahyar Pirzadeh Memorial Award for Outstanding Paper in Meshing Visualization and Computational Environments at the American Institute for Aeronautics and Astronautics (AIAA) Annual Science and Technology Forum this January. This paper discusses the process by which data can be transferred between disciplines in a way that increases accuracy while decreasing complexity.
New Funding from Air Force Research Lab at Wright Patterson Air Force Base
Through a $1.5 million award from the Air Force Research Lab, Dannenhoffer and Haimes are taking their analysis to the next level and are creating an open-source software solution that will streamline and accelerate the design and analysis process.
There are currently a number of tools on the marketplace that allow different types of engineering analysis to be completed. The problem is that these systems don’t easily integrate with one another. The tools and techniques that Dannenhoffer and Haimes are developing integrate a number of tools they have already designed, as well as systems that still have to be developed.
Some of the tools being integrated into the software are:
- Engineering Sketch Pad (ESP), a browser-based constructive solid CAD system
- OvrCad, a grid generation tool that generates overset grids directly from a parameterized solid model. It is a way to build CFD solutions very efficiently.
- Lumped Structure Model (LSM), a tool designed by MIT, that acts as a bridge between simple models and finite element analysis
- OpenCSM/EGADS, a geometry subsystem
- Analysis Interface and Meshing (AIM), which will be developed as part of the contract
- Legacy Geometry Converter, an interface that allows a user to generate a solid model from a cloud of points
- Computational Analysis Prototype Synthesis (CAPS)
Ultimately, these new tools and techniques and the accompanying software system will benefit companies and the government, which seek to find novel solutions to time-consuming and complex analysis and design problems.