Method of damage location determination based on a neural network using a single fiber Bragg grating sensor

In this paper, a structural damage identification algorithm based on a single fiber Bragg grating (FBG) sensor is proposed. The signal detected by the FBG can be analyzed by the wavelet packet decomposition and back propagation neural network to obtain the damage location information. A high-speed FBG demodulation system based on a tunable Fabry–Perot filter and unbalanced Mach–Zehnder (M-Z) interferometer is designed to respond to a signal with a frequency range from 0 to 4 kHz, which will increase the sensing accuracy. This algorithm is verified by the aluminum plate model, which can simulate the generation of damage in reality. The experimental results show that a single FBG sensor is enough to realize accurate damage location identification according to this algorithm, and identification accuracy can reach 90.0%.