Detail Design

The next phase of the project involved representation and location of all the major vehicular components as represented by 3D primitives. These primitives provided early fit checks of key components, their projected weight, and group properties like the vehicle’s center of gravity. More importantly, this critical engineering information enabled the team to make intelligent decisions early on in the design cycle, thus reducing the threat of error late in the manufacturing process.

Once the component primitives were selected from packaging constraint satisfaction and vehicle simulations, development of a detailed master representation began in earnest. Hundreds of parts were meticulously modeled, and the transformation from a simplified representation to a realisticlooking virtual prototype began. Almost no supplier could share geometric data on their products. As components were procured, detailed modeling and packaging often decided whether they stayed or were returned.

A six-month delay in the acquisition and measurement of the engine required late-process design flexibility when one of the engine dimensions exceeded even the generous primitive dimension assumed. Rapid flexing of the packaged assembly, however, led to an acceptable compromise without loss of performance or great impact on parallel design tasks. Immediate feedback led to an on-the-fly body design variation just as the body was being finished. Overall vehicle performance was reevaluated at successive levels of design complexity using our own custom vehicle simulation. Acceptable performance was maintained at 80 mpg and 0-60 in 7 seconds.

With all of the detailed subsystem modeling completed, the game of “what-if” concerning intelligent space claim began. Using component interference, collision information, and other geometric and mass property data, the team cyclically patched and refined the design. The final master representation file grew to more than 360MB and contained more than 10,000 features.

The final assembly layout came quickly. Using Pro/ENGINEER’s large assembly management solutions, the design lead could suppress unnecessary components, use envelope geometries, and explore different packaging arrangements quickly and efficiently. Concurrent design of mutually constrained subsystems was also facilitated. Team members worked in parallel on issues related to chassis design, transmission design, the routing of cables and piping, and ergonomics. Pro/ENGINEER’s assembly interchange functionality, design manager, and advanced utilities permitted evaluation of alternate configurations on the fly, and controlled the level of component detail during large assembly design review.

With the diverse analysis tools of Pro/MECHANICA, the design lead could determine a more optimal structural design for the chassis, find higher strength-to-weight ratio for the transmission’s gears, perform a motion analysis of the suspension assemblies, and determine safety factors in suspension components. Pro/MECHANICA also permitted the complex space frame structure to be quickly analyzed as simplified assembly of 1D beam primitives.

With this simplification and parametric associativity, successive design iterations could be performed from within the FEA environment. Easy-to-read design summaries were used to compare various design approaches. More than 40 FEA solution iterations over a few hours’ time recommended a final chassis design weighing just 186 pounds and satisfying a combined 3g bump load, a 2g braking load and a 1g cornering load. This optimization returned a 60% weight reduction and a 57% reduction in the number of different tube sizes compared with initial design concepts.

In addition to the chassis components, the unique design of our parallel format transmission was also analyzed. Gears were analyzed and changes made to reduce mass. This process was aided by Pro/MECHANICA’s parametric optimization capabilities. Performance measures such as mass, tooth root stress and deflection were used to determine optimal geometry for the gears. A kinematics analysis was performed on the rear suspension system to ensure proper motion paths and geometry changes for the A-arm design. The proper A-arm lengths, mounting points, and adjustment angle requirements were determined before any manufacturing began.

All these tasks were performed without physical mockup. Pro/MECHANICA’s powerful parametric optimization, when associatively coupled with Pro/ENGINEER, delivered a desktop capability to analyze, optimize and finalize a design in a seamless, integrated virtual environment.

Manufacturing

Building it right the first time was a requirement. The team’s expectations had now risen to take on the ultimate challenge: to build the vehicle that we had designed.

The transmission designer and design lead formulated a plan: learn and use Pro/MANUFACTURE ® . PTC’s manufacturing solutions provided the ability to rapidly create tool paths that represented the chosen geometry. Verification was performed with the enhanced simulation tools provided by Pro/NC ® . Once the tool paths were chosen and verified, they were then postprocessed and transferred to our HAAS VMC. After numerous trips to the milling center, 800 lbs. of billet 6061 aluminum became a scant 70 lbs. of shiny aluminum cases ready for assembly.

The completed transmission, the linchpin of our design, served as the base feature in the final drive assembly and enabled full electric drive operation of the L3 on January 29, 2001. The team and the car celebrated seven minutes of live television coverage and the resulting general stir in the community.

	Fig. 5. Norm Lamar driving the L3 during a televised event.

The Learning Process

We think the L3 may be the ultimate academic case study for modern design and lean manufacturing, and hope to use our experience to communicate our approach to other universities across the country and around the world.

First and foremost, we would like to emphasize the joy we feel from being able to exercise the freedom to actually make the product we designed—a product of quality that shows (not just tells of) our competence with the best tool set available, while at the same time honoring the memories of our friends.

Nevertheless, conducting this project in an academic setting had its challenges. Gaining sufficient mastery of the tools to undertake an ambitious project like the L3 requires a tremendous time commitment from faculty and students alike. At bottom, the problem is an academic reward structure that resists rewarding faculty for keeping current with broad interests unrelated to research and archival literature.

Premature demand for the product of a strong CAD/CAM/CAE infrastructure is an associated issue. Since investment is extensive and precedes noteworthy accomplishments like the L3 by several years, the administration may remove resources just as the payoff is materializing. The most important element of successful implementation, though, lies in educating the pioneering students about the value of PTC software skill sets. At first, only a handful of juniors could be convinced to invest themselves in the discipline of the tools. Of that initial group, some abandoned the tools because no educational synergy yet existed to reward their efforts across the curriculum. A very few went on to become our “poster children” with local industry— harbingers of the coming strength of our educational programs in this area. Still fewer enrolled as graduate students. These students provide the continuity of productivity to permit flagship projects—especially those with large information content—to succeed.

James S. Burns is Associate Professor of Mechanical Engineering at San Diego State University. Norman L. Lamar is an MSME student and a Design Engineer at Solar Turbines, Inc. in San Diego, California, USA. Jim can be reached by email at jburns@mail.sdsu.edu. More details on the L3 project are available at www.njaneer.com.