Abstract
An approach for robust design of mechanical systems using experimental design techniques is presented. Performance and achievement of functional requirements depend critically on variations on the geometry, dimensions, and positions of the system components. These variations arise due to limitations imposed by the capabilities of manufacturing processes, the tight control of which can be prohibitively expensive. This paper presents an approach for achieving robust designs which are less sensitive to these variations. Two formulations are used where manufacturing tolerances are considered as: (1) control variables; and (2) noise variables beyond the designer's control. Orthogonal arrays and analysis of variance are used to detect the most important design parameters to ensure performance robustness. The proposed design methodology is illustrated using a finger follower cam valve system. A comparison with other conventional design approaches is included.
Get full access to this article
View all access options for this article.
