| Early detection of structural damage has recently received much attention because of the promise that it could lead to economies in the maintenance of structure and provide warning of a later catastrophic structural collapse and the serious consequences that may follow. Researchers in the field have attempted to identify physical or vibration characteristics of the structure that could provide reliable indications of the health of a structure. The traditional methods of damage detection include visual inspection or instrumental evaluation. A comparatively recent development in the health monitoring of civil engineering structures is vibration-based damage detection. It offers several advantages. However, in practice there exist a number of challenges in vibration-based damage identification. In fact, most of the available damage identification algorithms fail when applied to practical civil engineering structures.; In this thesis, a new vibration based structural damage identification algorithm is proposed that may be able to overcome some of the difficulties. The proposed method combines structural dynamics and artificial neural network techniques. It splits the problem into two distinct sub-problems. In the first step, the damage location is determined by using the strain energy technique. In the second step the damage magnitude is evaluated by applying artificial neural network technique. It is expected that the proposed method will be both practical and effective. |
