Public Release: 

UB Researchers Develop First Computer Tool that Morphs Designs of Automotive and Aircraft Parts

University at Buffalo

BUFFALO, N.Y. -- Like the cartoon kids who zap themselves into Mighty Morphin' Power Rangers, engineers designing complex systems like cars and airplanes may now morph their design representations, thanks to graph morphing -- a powerful, new computer visualization method developed at the University at Buffalo.

The new method allows engineers to visualize how changes in their designs will affect design requirements.

The UB researchers also have developed a way to apply virtual reality to this new tool, allowing different individuals to access and modify the same three-dimensional design across a computer network or the World Wide Web.

Christina L. Bloebaum, Ph.D., UB associate professor of mechanical and aerospace engineering and principal investigator, said graph morphing will significantly improve the design process in any industry that produces a complex product, particularly the aircraft and automotive industries. It also will help improve the design and layout of industrial plants, where it will help planners to decide how changes in specific configurations would impact the function of the plant.

Bloebaum, who developed graph morphing with UB doctoral candidate Eliot Winer, will describe it on Wednesday, May 28, at the Second World Congress of Structural and Multidisciplinary Design Optimization in Zakopane, Poland.

She also will describe the first known application of Virtual Reality and Virtual Reality Modeling Language (VRML) to multidisciplinary design optimization, the field of engineering concerned with making complex, large-scale design more efficient and less costly.

Graph morphing allows designers to transcend the limits of two-dimensional computer screens by demonstrating how changing a variable in a multidimensional design space will affect the constraints and objectives that comprise the design representation of a complex system.

Until now, according to the UB researchers, designers working on such large-scale, complex design projects typically would be restricted to a very scaled-down visualization method that would address only two or three variables out of the entire problem.

"Before we developed this capability, it was virtually impossible to visually judge the impact of a change in a design variable in a complex, multidimensional design problem," said Bloebaum.

"The purpose of graph morphing is to capture in a two- or three-dimensional space on a computer screen both the objectives that are being designed for, such as minimizing cost, as well as the design constraints, such as maximum stresses or displacements or any of the other physical and performance limitations typically imposed in complex design," she said.

Working in the Multidisciplinary Optimization and Design Engineering Laboratory (MODEL) at UB, the researchers first developed graph morphing to apply to two-dimensional designs and then extended it to three-dimensional designs.

The objective is to examine potential trade-offs, situations where a change in one design parameter significantly affects one or more design objectives, constraints or disciplines -- such as structures or aerodynamics -- in a multidisciplinary environment.

"In a complex design environment, you have all these different subsystems that relate to each other in different ways, so when you change a variable over here, it changes what's happening elsewhere," Bloebaum explained. "It can be very difficult to understand those relationships without some sort of visualization capability."

In two-dimensional graph morphing, only two design variables may be plotted on the "x" and "y" axes; the others are assigned to, and controlled by, switches, which function like scroll bars on a screen. A designer may change one of these design variables by using a computer mouse to grab the switch that controls that variable.

The image will respond by almost immediately morphing to a visual representation of that change, as well as of the subsequent design changes that modification will cause, as the switch is moved again to new design variable values.

Bloebaum explained that virtual reality takes the power of graph morphing one step further.

She cited the example of a designer working on an aircraft wing who is not sure whether or not the span of the wing should be longer.

"With virtual reality, you can actually reach out and pull that wing longer and then see a representation of what will happen to the design space in three dimensions so that you can move around it and examine it from all angles," Bloebaum said. "It's easy to see the tradeoffs."

In addition to sheer aesthetics, graph morphing with virtual reality will represent in three dimensions the engineering impact of tradeoffs, such as how lengthening the wing would affect lift and drag and other critical variables in the final product's function.

With VRML, the UB researchers are providing a way for members of a design team to see how changes made by them and their colleagues, even those working in a different location, affect the performance of a product.

"When used with virtual reality, this technique will allow people working in different parts of a company in different cities, even on different types of computers, to work together simultaneously and to see how one modification in a design will affect other aspects of it," Bloebaum explained.

"We are using VRML to create a meaningful, three-dimensional design space that designers have access to across a computer network," said Bloebaum.

Last fall, Bloebaum developed the first method for designers with different computer platforms to utilize Java, the language of the World Wide Web, for communicating about multidisciplinary design optimization projects over the Web.

"Java now allows designers to work on the same project across heterogeneous computer environments," said Bloebaum, "and VRML is the virtual reality version of that."

The research is funded by Bloebaum's Presidential Faculty Fellow Award through the National Science Foundation.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.