HEALTH MONITORING OF A STRUCTURAL MEMBER

Abstract

I-lealth monitoring of structures has become an interesting research topic in recent years. To ensure the safety of the structure, crack detection at some early stage is very important. A study based on the vibration response of the cantilever beam is proposed in the present report. Health monitoring is done with the help of wavelet transformation of the vibration signal from the damaged beam. The methodology is based on ability of wavelet to detect minute abnormality in the vibration mode shape signal. In analytical study a finite element model is used and the mode shape is obtained with the help of a Matlab programme. In experimental demonstration the acceleration time graph are obtained from different accelerometers and then that signal is Fourier transformed. Mode shapes so obtained are then wavelet transformed. The wavelet used is the Haar wavelet (continuous wavelet transformation) in experimentally obtained signal and cgau-5 (a complex wavelet) in analytical study. The abrupt change in wavelet coefficients gives the presence of crack at that position. Vibration data from the intact structure is not required in proposed method. An optimization have been done for detecting maximum number of cracks, in case of multiple cracks of different severity in cantilever beam. The number of test points and the extent of crack are optimized from the huge collected data.