Abstract:
This dissertation compiles the development of an automated structural health monitoring
(SHM) for a cantilever structure. Three methodologies for the structural health monitoring
have been developed and validated. The first technique includes fiber optic sensors integrated
with the neural network method. This part of the research involves the collection of data
received from the fiber optic signal and integration of various optical parameters with neural
network for the strain determination. The strain is determined for both static and dynamic
loadings. Extension of work includes damage location in cantilever beam structure.
Second technique uses wavelet-based diffuse wave signals along with the neural network
method. In this section of the study, a wavelet-based diffuse zone selection method is
developed. The wavelet parameters are used in the developed neural network for damage
location in the structures.
The third methodology is purely wavelet-based diffuse wave analysis which can be used
structural health monitoring. This part of the research includes the development of a new
wavelet-based residual energy method for damage analysis. Results of the developed
methodology are compared with pre-existing techniques namely time-domain differencing
and spectrogram differencing and found in good agreement.
The study addresses issues related to smart structural health monitoring and proposes the
integrated approach for developing advanced measurement systems. It has also been shown
that integrated methods using artificial intelligence can be successfully used for better and
accurate structural health assessment.
