Multi-fault situation assessment and diagnosis method of wind turbine blades based on fault tree and Bayesian network

Abstract

A multi-fault diagnosis model of wind turbine blades based on a fault tree and Bayesian network was proposed to promptly obviate the frequent faults of wind turbine blades, prevent malignant accidents, and improve the reliability and safety of eolian power generation systems. Firstly, a fault tree model of wind turbine blades was established based on historical fault statistical data. The structural importance of basic events of the fault tree was subsequently acquired through qualitative analysis. According to the relationship between logic gates and basic events expressed in the fault tree structure model, a Bayesian network model was generated by a conversion algorithm, and a Bayesian network fault diagnosis model based on expert knowledge was further established. Finally, the prior probability of root nodes of the system failure was calculated through fuzzy comprehensive evaluation, and the posterior probability of each fault node was estimated according to the Bayes formula. The proposed method can diagnose and predict multi-fault modes of wind turbine blades and improve the troubleshooting efficiency of wind turbine blades. Moreover, the accuracy and feasibility of the proposed model have been verified by the case study, which indicates that the proposed method is of significant application value in engineering practice.