On the Minimum Required Sampling Frequency for Reliable Fatigue Lifetime Estimation in Structural Health Monitoring. How Much is Enough?

Structural Health Monitoring (SHM) represents the course of action of implementing a damage assessment strategy for engineering infrastructures. SHM systems can provide substantial aid towards the improvement of Offshore Wind turbines (OWTs) reliability, sustainability, and profitability. Usually, SHM system development is affected by three major concerns: the sensing technology, the associated signal analysis, and the interpretation algorithm. In this work, we focus on the relevance of the signal analysis on fatigue, being one of the most relevant damage sources. At some stage of the signal analysis process, analogue signals from strain transducers shall go digitized for computer analysis. In this phase, the engineer pursues the trade-off between gathering all the necessary information and storing the minimum data quantity. The sampling frequency adopted is paramount to this aim and can have substantial effects on the final lifetime estimated by damage accumulation rules. The Nyquist-Shannon sampling theorem is not well suited for minimum sampling frequency estimation in OWTs SHM. It allows recovering the frequency content of the original signal (bandwidth limited) yet does not guarantee correct reconstruction of its amplitudes (quantisation error). On top of that, the quantisation error is always on the non-conservative side of the lifetime estimation. We, therefore, provide examples showing that a ratio of the signal maximum “significant frequency” to the sampling frequency greater than or equal to ten (as opposed to two in the Nyquist-Shannon sampling theorem) is the rule of thumb to follow to avoid lifetime underestimation.