Statistical damage detection based on experimental acceleration resposnes

Wei Ming Li, Hong Ping Zhu, Lie Yun Ding, Yong Xia

Research output: Journal article publicationJournal articleAcademic researchpeer-review

1 Citation (Scopus)

Abstract

Four statistical damage detection methods were investigated as for five scenarios of a frame structure based on experimental acceleration responses. The frequency domain features were inspected. The distribution of occurrence rate of acceleration magnitudes in time domain was discussed, and the distributions of probability density and probability cumulative, as well as the trends of mean and standard deviation were analyzed. It is concluded that the changes of these features make the damage detection possible, but the method may be not feasible if considering the experimental error and the noises effect. The hypothesis tests were performed with a 5% significance level, and the damage detection was done based on the principal component analysis method. It is indicated that the results of hypothesis test can be used for damage existence detection, while the principal component analysis method can be used for both existence detection and preliminary quantification detection.
Original languageEnglish
Pages (from-to)107-110
Number of pages4
JournalZhendong yu Chongji/Journal of Vibration and Shock
Volume29
Issue number8
Publication statusPublished - 1 Aug 2010
Externally publishedYes

Keywords

  • Acceleration response
  • Damage detection
  • Frequency domain
  • Hypothesis test
  • Principal component analysis
  • Time domain

ASJC Scopus subject areas

  • Mechanical Engineering
  • Mechanics of Materials
  • Acoustics and Ultrasonics

Cite this