By Shyamal Roy
San Jose, Calif.
Edited by Paul Dvorak
Studies indicate that engineers nail down up to 80% of a product's final cost in the first 10% of a project. But when ideas are in flux and design goals are moving targets, CAD and FEA are not the most appropriate tools.
Hard facts are hard to come by early in a design. Nevertheless, this period is special because it's when mistakes are least expensive to correct. So it behooves the wise designer to test many ideas and make mistakes early on, before advancing to where corrections and changes get expensive. Software that provides more hard facts to the designer, earlier, could prove invaluable.
Predesign work often prompts questions such as, when will we know how sensitive belt tension is to tolerances in pulley diameter and belt length? How much work should it take to optimize an extrusion profile so the material extrudes straight? How long will it take to determine section properties for a complex beam? Or, why is it necessary to wait until finishing a design to check for interferences in the mechanism? The common thread through these questions is that design decisions are waiting on answers CAD programs cannot deliver.
Initial attempts at developing this so-called premodeling design-analysis (PMDA) software had limited success, primarily because they were built around variational-modeling techniques. Variational modeling involves spawning and solving a system of nonlinear simultaneous equations to track geometric constraints. While this method is flexible and can accommodate many situations, it slows as constraints are added. In addition, many of these programs wanted to occupy the entire design chain, thereby competing unsuccessfully with established CAD and solidmodeling packages.
An ideal conceptual program should help users visualize, analyze, simulate, and solve geometry-dependent challenges in the early stages of design. A premodeling design-analysis tool would be put to work while users have maximum flexibility, minimum information, and the most questions.
Users wishing to perform a PMDA need to start with a flexible model of the design intent. Then they can approach the challenge from a bottom-up or top-down view. In the bottom-up perspective, users capture the design intent in flexible models and then vary the inputs while receiving instantaneous decision support information. In the top-down approach, a user starts with an answer or requirement and the computer assists finding optimal design solutions.
Evaluating a set of givens at the start of a design may result in an optimal solution. Spreadsheets have been one way for testing early ideas. With some work, they can be rigged to drive a CAD system. But PMDA technology should go one better by letting users capture the functional intent of designs by associating geometry and calculations in a single expression. So as geometry and design variables change, the predesign software should recalculate and support instantaneous design decisions. In the following example, a PMDA program has captured parameters that contribute to a part's deflection.
PMDA software should also allow kinematic studies for mechanisms. Often, users can apply graphical solutions to solve complex geometry-based design challenges, all without solving a single equation. In the beltdrive example, as the user moves the driver pulley, belt tension and length automatically recalculate. This graphical analysis, called behavior modeling, finds solutions to design challenges that benefit from optimized shape, fit, and position. This geometric goal-seeking technology should let users start with results or required values and have the computer "backsolve" for exact geometry. For example, behavioral goals could be a required area, length, angle, perimeter, moment of inertia, volume, cost, weight, clearance, distance, tolerance, stress, or deflection. A heat sink provides another example. The software backsolves the geometry of the fins to fulfill a stated profile-perimeter target.
Many problems need only a concept sketch, a few key dimensions, and often only a few minutes of computer time. To be fair, more fluid ideas can take a few hours to solidify. Afterwards, designers can construct costly physical models and prototypes as necessary, and with more confidence that they'll work. Premodeling design analysis is an essential step that can deliver the maximum impact on the final design performance.