A novel methodology for health assessment of existing steel structures

Achintya Haldar, Rene Martinez-Flores

Research output: Contribution to journalArticlepeer-review

1 Scopus citations


A novel methodology is presented for health assessment of existing steel structures after natural and manmade hazards or due to natural aging process. The analytical concept and its desirable features are briefly discussed first. Then, the method is extensively verified using computer generated analytical response information and the response information obtained in laboratory investigations on defect-free and defective beams and a three-story one-bay steel frame. Two cracks were cut in the beam. Several defects were introduced in the frame in terms of removing a member, reducing the cross sectional area over a finite length, introducing several cracks and a single crack, presence of different defects at different locations, etc. The capability of the method was studied to check if it could identify the defects and their locations. The research team called the method as the GDLS-EKF-UI method. It is essentially a finite element-based system identification technique. It can identify defects at the local element level using only limited noise-contaminated response information and without using any information on excitation. The method is conclusively verified. It is recommended that the method be used to assess health of existing steel structures.

Original languageEnglish (US)
Pages (from-to)82-91
Number of pages10
JournalJournal of Structural Engineering (Madras)
Issue number1
StatePublished - Apr 2008


  • Experimental verifications
  • Finite element method
  • Safety assessment of existing steel structures
  • System identification
  • Time domain approach

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Mechanical Engineering


Dive into the research topics of 'A novel methodology for health assessment of existing steel structures'. Together they form a unique fingerprint.

Cite this