Multi-failure reliability evaluation of wind turbine blades under extreme conditions

Authored by

Xue Qin Li, Qiu Sheng Li, Matthias G.R. Faes, Michael Beer

Abstract

Wind turbine blades operating under extreme wind conditions are subjected to highly stochastic aerodynamic loading, which induces complex and strongly nonlinear structural responses and consequently introduces pronounced uncertainties into their structural behavior. To address these challenges, this study proposes an extreme-aware sparse Gaussian process regression with distributed collaborative (E-SGPR-DC) framework for system-level reliability analysis of wind turbine blades. The proposed framework combines a distributed collaborative strategy to effectively combine local sub-model information with global response characteristics and embeds an extreme-aware learning mechanism to enhance prediction fidelity in high-stress and near-failure regions. By further integrating probabilistic strength characterization and Gaussian Copula-based dependency modeling, a unified multi-failure reliability evaluation framework is established. Numerical investigations demonstrate that the proposed approach significantly improves prediction accuracy and computational efficiency compared with conventional surrogate methods, while providing more reliable estimates of structural reliability under correlated failure modes. These results indicate that the E-SGPR-DC framework offers a robust and efficient tool for structural reliability evaluation and design of wind turbine blades under realistic and extreme operational uncertainties.

Details

Organisation(s): Institute for Risk and Reliability
External Organisation(s): City University of Hong Kong
TU Dortmund University
Tongji University
University of Liverpool
Type: Article
Journal: Structures
Volume: 88
ISSN: 2352-0124
Publication date: 06.2026
Publication status: Published
Peer reviewed: Yes
ASJC Scopus subject areas: Architecture, Civil and Structural Engineering, Building and Construction, Safety, Risk, Reliability and Quality
Electronic version(s): https://doi.org/10.1016/j.istruc.2026.111836 (Access: Closed )