特徵辨識的用途之一在於協助CAE分析之前置處理，事先辨識出特徵所在，方能針對不同特徵進行網格加密或簡化的動作，以便建構出合適的網格，而幾何特徵(Hole、Fillet、Rib、Chamfer等)是影響網格品質的重要因素之一，本研究將以孔洞特徵為主要辨識對象，加入混接面辨識技術，發展出過去鮮少針對具有混接面的孔洞特徵辨識技術，同時將孔洞的外形進行有系統的歸類，記錄孔洞相關資訊。本研究提出一基於B-rep模型之孔洞特徵辨識技術，期望以自動化取代傳統以肉眼辨識的方式，大幅降低數值分析上所需花費的時間。此外，將以多組實際案例來驗證孔洞辨識技術的準確性與可行性，驗證本研究的可信度。;The purpose of feature recognition is for the pre-processing of the CAE analysis. In the beginning of the analysis, various types of features must be recognized, in order to it can generate meshes of high density. The feature recognition can also be used to simplify the features, so as to build high quality meshes. Some geometric features, such as holes, fillets, ribs and chamfers are one of the key factors affecting the quality and accuracy in finite element analysis. This study focuses on the recognition of various kinds of holes. Particularly, a blend face recognition algorithm is added to deal with the cases with fillets at the boundary. Meanwhile, various kinds of holes are classified and the associated data for each of them are recorded. This research presents an approach based on the B-rep model for the recognition of holes. It can replace traditional manual work significantly, and hence improve the efficiency of meshes generation in the CAE analysis. Several realistic CAD models are employed to verify the feasibility of the proposed algorithm.