1999 Fiscal Year Final Research Report Summary
SYSTEM IDENTIFICATION OF STRUCTURES BY WAVELET TRANSFORM
| Project/Area Number |
10650242
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Dynamics/Control
|
|---|
| Research Institution | KYOTO INSTITUTE OF TECHNOLOGY |
|---|
Principal Investigator |
SONE Akira FACULTY OF ENGINEERING AND DESIGN, KYOTO INSTITUTE OF TECHNOLOGY, PROFESSOR, 工芸学部, 教授 (20197015)
|
|---|
| Project Period (FY) |
1998 – 1999
|
| Keywords | IDENTIFICATION / EXPERIMENT / WAVELET TRANSFORM / STRUCTRAL PARAMETER / MEXICAN HAT |
|---|
| Research Abstract |
In this research, a system identification technique for structures is proposed when the structures are damaged due to external disturbance such as earthquake. In order to detect the change of structural parameters like mass, stiffness and damping of structure, parametric system identification technique using wavelet analysis which provides the time-frequency analysis, is proposed and its validity is comformed through the numerical simulations and experiments. This method has the advantage such that without measuring acceleration response and veleocity response, the structural parameters are estimated by using only the displacement response. Furthermore, since the information of signal about time is obtained by the wavelet transform, the time-varying structural parameters can be identified. The results obtained from this research are summarized as follows.
(1) From the results obtained using two-degree-of freedom structure, five-degree-of freedom structure and simply supported beam modelled by finite elements, it is clear that the wavelet transform can effectively estimate the structural parameters.
(2) It has been proven that the change in stiffness and daping of structures can be identified by the proposed technique.
(3) The experimental results using one-story structure demonstrate that the proposed technique is viable for the identification of stiffness. We do not obtain good results compared with half power method in low damping case.
(4) Since the accuracy of identification depends on the sampling time for data, it is clear that the adequate samping time is required in order to overecome the disadvantage of wavalete transform.
|
