Optical fiber sensing under high strain gradients: a multi-technology benchmark against finite element predictions
Abstract
Distributed and quasi-distributed optical fiber sensors (OFS) are deployed for structural health monitoring (SHM) of structures, yet their accuracy near sub-centimetre stress concentration areas remains uncertain. Three centimetre-scale sensing technologies are benchmarked against finite element method (FEM) predictions: fiber Bragg gratings (FBGs) in wavelength-division multiplexing (WDM), continuous fiber gratings (CFGs) in frequency-modulated continuous-wave (FMCW), and optical frequency-domain reflectometry (OFDR) under high strain gradients (HSGs) reaching around 200 µε mm−1. HSGs are generated in a 2 mm thick aluminium plate (80 × 765 mm2) with three holes (diameters Φ = 30 mm, 40 mm and 50 mm) loaded up to 2200 N by steps of 200 N. Gage-length averaging is the dominant error mechanism at the 40 mm and 30 mm holes when compared with the FEM structural benchmark: an 8 mm type-I FBG underestimates the FEM peak by 24 % and 30 %, respectively, while a 3 mm type-III FBG reduces this to around 10 % to 15 %. At the 50 mm hole, FBG type-III remains below FEM by about 12 % (1941 µε vs 2209 µε at 2000 N); at 1000 N, FEM, OFDR, and FMCW peaks are 1113.4 µε, 1332.25 µε (+20 %), and 555.4 µε (-50 %), respectively. OFDR requires moving-reference processing above 1600 N to prevent correlation breakdown. The relative accuracy of the 2 mm and 9 mm CFG-FMCW configurations varies with hole size and depends on gage overlap fraction; FMCW damping or spectral distortion occurs in high-gradient regions and is consistent with gradient-dependent spectral distortion, potentially involving chirp and proprietary peak-detection processing. These results provide guidance for sensor selection, gage-length specification, and signal-processing choices in SHM deployments where damage-induced strain concentrations are expected. Mitigation perspectives include transfer-function-informed correction (regularized deconvolution) and hybrid OFDR–FBG data fusion to preserve peak fidelity while retaining spatial coverage.