Generative Process Planning of Prismatic Parts by Feature Relaxation
Abstract Feature-based product models have recently been proposed as a basis for generative process planning systems for mechanical applications. In this approach, the part is broken into a set of manufacturing features which are associated with various kinds of technological information useful for process planning. A fundamental problem of this approach is the fact that a given part usually has several interpretations as features. Ideally, all these interpretations should be taken into account in the process planner in order to achieve globally optimal plans. Hence, any planner that starts from a fixed collection of features created by feature recognition or by user input has already committed itself to a limited view of the part, and cannot take into account manufacturing opportunities corresponding with the other views. As a solution to this problem of premature commitment, we propose the use of what we call relaxed feature models. In this approach, features can be reinterpreted by the process planner to take into account manufacturing possibilities from a wider range than what any particular selection would make possible. As an example of the benefits of this approach, we describe the manufacturability analysis component of our experimental generative process planner for 3-axis milling operations of prismatic parts.