卷积神经网络
领域(数学)
学习迁移
深度学习
任务(项目管理)
计算机科学
人工智能
拉伤
多任务学习
结构工程
拉伸应变
机器学习
极限抗拉强度
材料科学
工程类
复合材料
数学
医学
内科学
系统工程
纯数学
作者
Zekai Huang,Dongdong Chang,Xiaofa Yang,Hong Zuo
标识
DOI:10.1016/j.engfracmech.2023.109703
摘要
This paper proposes a novel approach that utilizes a deep convolutional neural network (DCNN) for crack damage detection in thin plates. The DCNN model converts the detection task into a regression task and identifies crack tips through the strain field. Numerical simulations and experiments under quasi-static tensile were conducted to demonstrate the proposed method. The results indicate that this method exhibits high accuracy in crack damage detection. Furthermore, the study explores the use of active learning to address the challenge of data scarcity and extends the application of the DCNN model to similar tasks by transfer learning. This study provides some new perspectives for the field of damage detection.
科研通智能强力驱动
Strongly Powered by AbleSci AI