点击次数：

论文类型：期刊论文

发表时间：2019-01-01

卷号：101

期号：1-4

页面范围：963-977

ISSN号：0268-3768

摘要：Undesirable machining damages exist frequently in the conventional machining processes of CFRP, causing significant decline in the performance of CFRP components. For effectively suppressing various machining damages, this study aims to reveal the evolution laws of the fiber-matrix interface crack, which can be regarded as the origin of majority of the machining damages. In order to accurately and conveniently obtain the evolution laws determined by numerous coupled factors, the study is conducted through a combined method of experiments and machine-learning predictions. Specifically, orthogonal cutting experiments are conducted for analyzing the fiber-matrix interface crack's types and depth, and upon these characteristics predictive models based on machine learning are established for acquiring sufficient data so that reliable evolution laws can be obtained. From the results, it is found that there are four evolution types in all and when burrs and cracks exist, the crack depth increases obviously, and both the evolution types and depth are mainly determined by the fiber cutting angle. According to this, preliminary optimization methods of machining CFRP are proposed. The conclusions of this study are beneficial to facilitate the development of high-quality and high-efficiency machining technologies of CFRP.

DOI码：10.1007/s00170-018-2992-8