Implementing fuzzy-based artificial intelligence approach for location of damage in structures
Modal parameters are functions of the physical characteristics of a structure and they are very sensitive to damage. Therefore, any alterations in the physical features can change the vibration parameters of a structure. Modal data such as natural frequencies and mode shapes are easy to acq...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Horizon Research Publishing Corporation
2022
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/7569/ http://eprints.uthm.edu.my/7569/1/J14335_00d3ff4db49b7731feefe285912ac5be.pdf |
| Summary: | Modal parameters are functions of the
physical characteristics of a structure and they are very
sensitive to damage. Therefore, any alterations in the
physical features can change the vibration parameters of a
structure. Modal data such as natural frequencies and mode
shapes are easy to acquire from the measurements of
structural behavior. One method of structural damage
identification is to apply natural frequency. Natural
frequencies represent the global behaviors of a structure
and are not too sensitive when detecting the damage in
structures and cannot offer spatial information about
structural changes, and thus, their application is considered
as challenging. On the other hand, a mode shape is a
vibrational deformation of a system and it represents the
relative displacement of all parts of a structure and can
provide spatial information as well as give a significant
indication of the damage occurring in a structure. In this
present research, an intelligent hybrid approach, namely
adaptive neuro-fuzzy inference system (ANFIS), as a
fuzzy-based artificial intelligence approach was developed
and applied due to its ability to recognize patterns, strong
computational features, and capability of locating defects
in a scaled girder bridge using direct modal parameters.
The experimental analysis and numerical simulations of a
steel girder bridge provided mode shape parameter datasets
under different positions and sizes of faults in the structure.
The results demonstrated the effectiveness of this method
and provided acceptable precision even when the input
datasets contained errors or were corrupted with a certain
level of noise. |
|---|