Prediction of wind turbine blade icing fault based on selective deep ensemble model

作者：

Highlights：

• A selective deep ensemble model is constructed for the prediction of blade icing.

• Cost-sensitive learning is introduced to address imbalanced wind turbine data.

• Group method of data handling (GMDH) is used to realize selective deep ensemble.

• Our model outperforms existing ensemble models and single deep learning models.

摘要

•A selective deep ensemble model is constructed for the prediction of blade icing.•Cost-sensitive learning is introduced to address imbalanced wind turbine data.•Group method of data handling (GMDH) is used to realize selective deep ensemble.•Our model outperforms existing ensemble models and single deep learning models.

论文关键词：Wind turbine,Blade icing fault prediction,Selective deep ensemble,Imbalanced SCADA data,GMDH

论文评审过程：Received 16 July 2021, Revised 22 January 2022, Accepted 22 January 2022, Available online 31 January 2022, Version of Record 21 February 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108290