NC process reuse oriented effective subpart retrieval approach of 3D CAD models

• Published in: Computers in industry Vol. 90; pp. 1 - 16
• Main Authors: Xu, Changhong, Zhang, Shusheng, Huang, Bo, Li, Xiuling, Huang, Rui
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2017
• Subjects: Medial axis transform; NC process reuse; Similarity; Subpart retrieval; Subpart retrieval; Medial axis transform; NC process reuse; Similarity
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/j.compind.2017.04.006

•The association mechanism between medial axis transform (MAT) and NC machining is established to reveal the association between geometry and NC process.•The machining precision and cutting tools filtering for accelerating subpart retrieval is presented to filter out infeasible subparts quickly.•A multilevel feature descriptor capturing various levels of NC process information for reuse is designed.•The matched feature pairs similarity and feature layout information similarity are presented to calculate the similarity between subparts. As a huge number of 3D CAD models with NC process are generated each year, retrieval of 3D CAD models for achieving NC process reuse is becoming an effective strategy to improve the efficiency of NC programing. However, there has been little research on how to discover similar subparts for NC process reuse. In this paper, a NC process reuse oriented effective subpart retrieval approach of 3D CAD models is proposed. Firstly, the medial axis transform (MAT) of a pocket is introduced and adopted as the reference tool paths. The association mechanism between MAT and NC machining is established to reveal the association between geometry and NC process for features. Then the rapid filtering for accelerating subpart retrieval is presented to filter out infeasible subparts quickly under the requirements of NC process reuse. Moreover, a multilevel feature descriptor capturing different levels of NC process information for reuse is elaborated to construct the feature similarity assessment model. Finally, the matched feature pairs similarity and feature layout information similarity are presented to calculate the similarity between subparts. In the experiment, a prototype based on CATIA is developed and the experimental results are analyzed at length to verify the effectiveness of the proposed approach.