Feature recognition and machine learning in finite element models through a clustering algorithm

Abstract

The feature identification of the CAD model is a significant task in any CAD algorithm. Enormous computational time, huge memory allocation, lack of understanding in computational geometry, etc., are some of the complications faced while implementing the feature recognition algorithms. This paper represents a clustering algorithm procedure in finite element models, which is the predominant component in analysis methodology. This study performs the clustering of data groups through density-based clustering algorithms such as mean shift clustering and K-means clustering algorithm. In addition to that, experimental evaluation based on the structured algorithm procedure for identifying the features of CAD geometries is investigated. Finally, the study evaluates the performance of the proposed structured algorithm and its efficiency in terms of both computational time and computational memory.