Open AccessResearch on Ultrasonic Detection of Air Spring Rubber Debonding based on CEEMDAN
Author(s) -
Dawei Hou,
Xuemei Wang,
NI Wen-bo
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1549/3/032154
Subject(s) - natural rubber , ultrasonic sensor , materials science , vibration , hilbert–huang transform , acoustics , spring (device) , approximate entropy , composite material , computer science , structural engineering , pattern recognition (psychology) , artificial intelligence , engineering , physics , telecommunications , white noise
The SYS510e air spring of electric multiple unit (EMU) is the metal and rubber bonding structure. In order to carry out ultrasonic debonding detection on the tapered bonding interface, after correcting the interference caused by sectional thickness change, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and cross approximate entropy (ApEn) method was applied to analyze time frequency characteristics of echo under different bonding conditions. Five numbers of cross ApEn features were extracted to identify rubber debonding defect. These features are more stable and show more significant discrepancies between bonding and debonding states. Then BP neural network was trained to identify the debonding state of the tapered interface of air spring. The results of ultrasonic C scan show that the ultrasonic detecting method based on CEEMDAN and cross ApEn can accurately and effectively identify the location and contour of rubber debonding defects, and meet the needs of the debonding detection of SYS510e air spring of EMU.