Local damage identification of offshore wind turbine towers using a squared strain-mode relative-change indicator: numerical and experimental validation

This paper proposes a squared strain-mode relative-change (SSMR) indicator for local damage identification in offshore wind turbine towers. By combining damaged-state normalization with squared enhancement, SSMR amplifies damage-induced strain-mode perturbations while suppressing abnormal amplification caused by small reference components. An 8.5 MW tower finite element model investigates axial cracks, tower-root circumferential fatigue cracks, and segmented-flange bolt fractures, while comparing the sensitivities of natural frequencies, displacement modes, and strain modes. Random hammer-impact tests on a thin-walled cylindrical specimen are further conducted, with strain modal parameters identified using covariance-driven stochastic subspace identification (SSI-COV). Results show that strain modes are more sensitive to local damage than natural frequencies and displacement modes. SSMR produces clear peaks at actual damage locations, retains good discrimination under multiple-damage conditions, and reaches a 67.26% damage-location indicator proportion at the crack location in the most severe single-crack case with a 12-point layout.