Online Fatigue Damage Detection for Structures under Random Loading Using a Comparative Sensor-Data Approach

Abstract

Fatigue failure of metallic structures such as wind turbines towers or steel bridges is a problem that affects their remaining service life. The current industry standard of fatigue damage assessment is based on time domain cycle counting to estimate the cumulative fatigue damage which is compared to a reference value to detect fatigue failure. The stochastic nature of the fatigue process and the uncertainty in choosing the appropriate reference value limit the applicability of this approach. Fatigue crack detection is the focus of many structural health monitoring techniques. A novel sensor-data comparative approach for fatigue damage detection is proposed in this paper. This approach is based on online monitoring of the linear correlation between monitored and reference sensordata. Fatigue failure is detected when the linear correlation is lost. The proposed concept is validated experimentally and the obtained experimental results prove the possibility of early detection of fatigue failure.