Computer models of complex mechanical systems are being used more frequently and effectively due to ever increasing computing power. These models have improved the design process by allowing designers to create virtual prototypes of systems. Computer models permit iterative designs, thereby allowing the designer to run hundreds of simulations in the time it would take to make one physical model and test it. This ease of iteration has made design of experiments (DOE) analyses easy to carry out on the computer models. A DOE analysis provides excellent information regarding the specific output function being examined but may overlook some of the global effects that these changes may have on the system. A new design tool is developed in this work that addresses this deficiency. The new method greatly improves upon existing design tools such as DOE. Specifically, the new method augments a traditional DOE study with an Activity study, which focuses on the energy flow in the system and provides a larger picture of the system dynamics. Instead of using only a DOE to determine which elements’ properties should be changed, the Activity of the elements is also considered. This helps prevent altering a parameter or set of parameters to achieve a desired response at the expense of other unforeseen results. This idea is demonstrated by using an ADAMS model of a small production vehicle to design for improved ride. The example shows that design changes based on a DOE analysis alone can improve the ride quality, but also can alter other system responses significantly. When an Activity analysis is combined with the DOE study, the ride quality is improved, but the effects on other aspects of the system are greatly reduced. Augmentation of DOE with Activity allows a specific design goal to be achieved and also provides global information about the system. This allows the engineer to make informed decisions about the effects of changes on the system.

This content is only available via PDF.
You do not currently have access to this content.