Recognizing machining features through partial constraint satisfaction

Machining features are groupings of geometric and topological entities that have certain engineering significance. Recognizing these features from a designed parts is crucial in the automation of computer integrated manufacturing (CIM). In this paper, the machining feature recognition problem is formulated as a partial constraint satisfaction problem (PCSP) where variables are the faces in the delta volume (the difference between a part and a stock), the possible feature classifications are values for the variables, and the constraints are the geometric and topological properties between the variables. Based on this framework, several techniques can be applied to solve the PCSP. An integral method that combines extended forward checking, variable ordering and value ordering is employed to solve the problem. Interacting features are then recognized by solving the PCSP and verifying the solution that partially satisfy all the constraints.