Constraint management techniques for the design of large-scale and complex mechanical systems

Engineering design is an iterative process involving the formulation and satisfaction of a network of constraints defining the design system. Although the solution strategy to handle these constraints varies from one design scenario to another, the underlying concepts of constraint management can be applied in a generic manner to any design phenomenon that can be represented by a mathematical model. A previous work used these concepts to develop an interactive framework called the design shell to facilitate the design of mechanical systems. The goal of the current research is to expand the scope and performance of this framework to handle larger constraint networks and more complex design specifications. The main constraint management issues considered here are design decomposition and design solution.;A two-stage design decomposition methodology for handling large constraint systems has been implemented taking certain activity information into consideration. Since routine design systems can involve all manner of constraints, a strategy for handling dynamic constraint systems using suitable constraint management techniques has been proposed. Directed graphs and graph matching algorithms have been used to develop methodologies for diagnosing and decomposing an under-constrained design system in the presence of dynamic constraints. These algorithms are also efficient both in terms of storage and speed since they exploit the inherent sparseness of typical mechanical design systems. The globally convergent nature of homotopy continuation methods is extremely viable for handling constraint networks for which estimates of the final solution is not available. Strategies involving bounded homotopy maps have been developed to determine more than one root in a prescribed domain. In addition, the design of a library of graphical interface routines for the development of intuitive and user friendly CAD applications has been studied and implemented.;The research also presents the development of comprehensive mathematical models for the design of helical springs. Complex design considerations such as material selection, fatigue analysis, and quasi-static constraint systems have been handled within an existing interactive design framework.