钢丝绳
电磁线圈
绳子
人工神经网络
材料科学
磁通量
无损检测
曲面(拓扑)
声学
漏磁
结构工程
计算机科学
工程类
磁场
人工智能
电气工程
物理
数学
几何学
量子力学
作者
Donglai Zhang,Enchao Zhang,Donglai Zhang
标识
DOI:10.1016/j.ndteint.2021.102405
摘要
The detection of internal and external defects in steel wire rope is a very important task. Current nondestructive testing methods cannot distinguish between internal and surface defects, and cannot determine the quantitative characteristics of internal defects. This paper first analyzes the advantages and disadvantages of induction coil magnetic flux detection and Hall sensor-based magnetic flux leakage detection. The problem of simultaneous detection using these two methods is then solved, and a new quantitative detection method is proposed for internal and surface defects in wire rope. Finally, a calculation method that uses a two-step training algorithm to establish two sets of neural networks is proposed. Results from simulations and experiments verify that the proposed method can accurately distinguish between internal and surface defects in wire rope, and can also quantitatively detect the width, cross-sectional loss rate, and depth of the defects.
科研通智能强力驱动
Strongly Powered by AbleSci AI