Finite element-based evaluation of minimum detectable damage in the Magerholm Linkspan under environmental, noise, and damage asymmetry effects

Structural health monitoring (SHM) of civil infrastructure is often challenged by environmental and operational variability, which can mask or imitate damage-induced changes in modal properties. This study evaluates the minimum detectable damage (MDD) of the Magerholm linkspan using a finite element—based sensitivity framework that explicitly accounts for environmental effects and measurement noise. Environmental variability was modeled using a synthetic temperature profile and a two-stage tidal stiffness interpolation scheme derived based on field-identified natural frequency data. Damage was simulated as a reduction in the stiffness provided by one or both of the lifting towers. The detectability of damage was quantified using the probability of detection and receiver operating characteristic curves derived from Mahalanobis distance-based damage indices. Results show that tidal variations exert the strongest environmental effect on natural frequencies, yielding a significantly higher MDD compared to temperature alone. When both effects were combined, MDD increased further, and the introduction of noise to the extracted frequencies caused a substantial reduction in detectability, with MDD rising from 22% in the noiseless case to over 35% with 3% noise. Under asymmetric damage conditions affecting a single tower, the frequency changes are more subtle, leading to reduced detectability. Moreover, the MDD increases sharply with noise to above 50% at 3% noise—highlighting the strong noise sensitivity of asymmetric configurations. However, detectability improved when modes with low stiffness sensitivity were excluded. Overall, the study demonstrates the strong impact of environmental variability and noise on the detectability of damage in linkspans and provides practical insights into achievable MDD thresholds for lifting towers using vibration-based SHM.