Failure analysis of floating offshore wind turbines considering working condition correlations

Abstract

AbstractA novel failure analysis method named D‐vine copula Bayesian Network is proposed, aimed to extract fault correlation information from various condition monitoring channels of floating offshore wind turbine components, quantify risk probability and consequence, and obtain risk priority of components considering condition correlation. First, a copula Bayesian model is established based on the Bayesian network and D‐vine copula theory. Then, a risk consequence calculation method considering relative loss is developed. Finally, critical failure items are identified by a risk matrix. The proposed technique is expected to: (i) Release the limitation that parent nodes with condition monitoring input are processed as independent in Bayesian analysis. (ii) Provide an alternative way for presenting relationships between nodes instead of conditional probability tables. (iii) Simplify the calculation of high‐dimensional copula. This study screened out high‐risk level components of floating offshore wind turbines, and operation recommendations avoiding potential failure risk are put forward. The comparative results demonstrate the feasibility and reliability of the proposed method.