Structural damage detection by improved back-propagation neural networks based on modal test data | |
| Author | Baisen, He |
| Call Number | AIT Thesis no. ST-01-14 |
| Subject(s) | Neural networks (Computer science) Structural dynamics |
| Note | A thesis submitted in patiial fulfillment of the requirements for the degree of Master of Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | Neural Networks (NNs) have played a very important pole in structure damage detection. However, because of the non-linearity of sigmoid transfer function, and the inappropriate initialized weights or other reasons, back-propagation method, the mainly used method in this area, encounters two main problems in practice: (1) The convergence tends to be extremely slow, (2) There are many local minina. In this research, two kinds of neural network algorithms (adding momentum terms method and rapid BP method) were adopted to improve the traditional Back-propagation (BP) neural networks. In each method, linear, tansig and logsig transfer functions were used and compared. Analytical and numerical results showed that both the proposed two methods are more efficient than the traditional methods and previous BP methods on structural damage detection. These trained neural networks can have the capability of recognizing the location and the extent of individual member damage from the measured modal data of the structure. Overall, the improved BP networks are reliable and facile methods in structural damage detection. |
| Year | 2001 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Structural Engineering (STE) /Former Name = Structural Engineering and Construction (ST) |
| Chairperson(s) | Zhu, Hongping |
| Examination Committee(s) | Pennung Warnitchai ;Takewaka, Koji |
| Scholarship Donor(s) | Government of Japan |
| Degree | Thesis (M.Eng.) - Asian Institute of Technology, 2001 |