A novel bi-component structural health monitoring strategy for deriving global models of operational wind turbines

Abstract

The short and long-term variability characterizing operational Wind Turbine (WT) structures limits applicability of existing Structural Health Monitoring (SHM) strategies for diagnostics and condition assessment. In this paper, a novel modeling approach is proposed delivering global models able to account for a wide range of operational conditions of a WT System. The approach relies on the merging of environmental and operational variables into the modeling of monitored vibration response via a two-step methodology: a) implementation of a Smoothness Priors Time Varying Autoregressive Moving Average (SP-TARMA) method for modeling the non-stationary response, and b) implementation of a Polynomial Chaos Expansion (PCE) probabilistic model for modeling the response uncertainty. The bicomponent tool is applied on long-term data, collected as part of a continuous monitoring campaign on a real operating WT structure located in Dortmund, Germany. The delivered statistical model of the structure yields a robust representation of the underlying structural dynamics, distinguishing actual structural damage from performance shifts attributed to environmental and operational agents.