Efficient residuals pre-processing for diagnosing multi-class faults in a doubly fed induction generator, under missing data scenarios
作者：
Highlights：
• Extended fault diagnosis system for a doubly fed induction generator.
• Improved the ensemble based decision module to allow incremental learning of new fault classes.
• The pre-processing module generates the latent residuals.
• The Wold cross-validation algorithm estimates the number of latent residuals.
• The scheme can diagnose the faults under missing data scenarios.
摘要
•Extended fault diagnosis system for a doubly fed induction generator.•Improved the ensemble based decision module to allow incremental learning of new fault classes.•The pre-processing module generates the latent residuals.•The Wold cross-validation algorithm estimates the number of latent residuals.•The scheme can diagnose the faults under missing data scenarios.
论文关键词：Fault diagnosis,NIPALS,Wold cross-validation,Latent residuals,New class faults,Wind turbine
论文评审过程：Available online 18 April 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.03.056
