关键词: 加工特征, 特征识别, 属性邻接图

Abstract: To extract 2.5 dimension machining feature from solid boundary model effectively, an available and feasible feature recognition approach was proposed. Referring to the actual machining procedure of the rough, the approach was composed of a series of steps, which included the ready to be machined volumes decomposition from the rough solid model, local composition feature recognition and extraction along the tool-feeding direction by tier, the generic feature composition was formed by redecomposing and recomposing the local composition feature according to the concave and convex of top face,and the feature relationship tree was constructed and created by matching relationships among features. The related algorithm to every step was given. These algorithms were based on graph-based Boolean operation, and the volume features were created by reconfiguring the feature boundary to find rounded feature boundary. This approach was suit to the integration with downstream units, such as Computer Aided Manufacturing(CAM).It was tested through typical 2.5 dimension machining part to illuminate the validity and practicability of the method.

Key words: machining feature, feature recognition, attribute adjacent graph, volume decomposition