Statistic model for calculating the hole-edge crack length using fiber bragg grating sensors

Fiber Bragg grating has been reported to be a potential tool in structural health monitoring (SHM). However, the application of this technology mainly focus on the composite health monitoring, seldom fatigue crack propagation researches on the aluminum material. This paper presents a new crack length quantification method based on fiber Bragg grating sensors (FBGs). The damage sensitive features were collected by characteristics algorithm. In this paper, the primary wavelength of the reflection intensity spectra and full width at half width (FWHM) were analyzed versus the crack length. The relationship between the crack length and damage sensitive characteristics was analyzed by the BP neural network, and 15 hidden layers were used in this model. The results of BP neural network were validated using coupon test data with fatigue crack propagation.