Presentation Details
  1. Home
  2. Presentation List
  3. Presentation Details
Data-Driven Algorithm-Based Blade Icing Fault Prediction for Wind Turbine
ID：34 View Protection：ATTENDEE Updated Time：2025-11-10 11:04:39 Hits：178 Oral Presentation

Start Time：2025-11-22 16:20（Asia/Shanghai）

Duration：10min

Session：S3 Parallel Session 3 » S3-1Parallel Session 3-22 PM

No files

Abstract
Wind turbine blade icing seriously threatens power generation efficiency and operational safety in cold regions. To address the problem of insufficient accuracy and interpretability in existing icing detection methods, this study proposes a data-driven blade icing fault prediction approach based on Supervisory Control and Data Acquisition (SCADA) operational data. The Synthetic Minority Oversampling Technique (SMOTE) is applied to alleviate class imbalance in icing samples, while Recursive Feature Elimination (RFE) is employed for feature dimension reduction and optimization. Based on the processed dataset, multiple machine learning algorithms—K-Nearest Neighbors (KNN), Random Forest (RF), Stochastic Gradient Descent (SGD), and XGBoost—are systematically evaluated. Results show that the XGBoost model achieves the best prediction performance, with an accuracy of 99.80% and a recall of 99.41%, effectively identifying icing events. To enhance model interpretability, the SHAP method is introduced to analyze the contribution of key features, revealing that wind speed, temperature, and humidity are the most influential factors. This study provides a reliable and interpretable data-driven framework for wind turbine blade icing fault prediction, offering technical support for intelligent operation and maintenance.
Keywords
wind turbine blade icing; feature engineering; data-driven algorithms; SHAP; imbalanced data
Speaker
Liuxu Wang
PhD graduate student Xi’an Jiaotong University

Submission Author
Liuxu Wang Xi’an Jiaotong University
Wei Deng Xi’an Thermal Power Research Institute Co. Ltd
Fang Wan china；Huaneng Gansu Energy Development Co., Ltd
Min Zhang Northwest University；china
shouwang zhao Xi’an Jiaotong University；College of Electrical Engineering Xi’an Jiaotong University
Ruolan Hu china；Xi’an Thermal Power Research Institute Co. Ltd y Xi’an, China
Gangli Fang Xi’an Thermal Power Research Institute Co. Ltd；china
Yong Zhao Xi’an Thermal Power Research Institute Co. Ltd
Yu Chen Xi'an Jiaotong University
Submit Comment
Verify Code Change Another
All Comments
Important Date

  • Conference Date

    Nov 21

    2025

    to

    Nov 23

    2025

  • Oct 20 2025

    Draft paper submission deadline

  • Dec 08 2025

    Registration deadline

Sponsored By
IEEE Instrumentation and Measurement Society
South China University of Technology
Organized By
South China University of Technology
Contact Information
Login to view full contact details
Previous Conferences