Abstract

In this paper, we demonstrated the use of the k-Nearest Neighbor, a machine learning algorithm, to identify mechanical vibration modes of a cantilever beam in a frequency range between 40-300 Hz at an accelerations of 1.1 ± 0.1 g. We attached fiber Bragg gratings to the cantilever structure and analyzed the spectral response during vibration. We observe small increases in spectral bandwidth of three Bragg gratings to perform a 3-dimensional classification environment and evaluated the accuracy of the algorithm with independent testing data.

© 2020 The Author(s)