Scholarlink

<p>Wind energy represents a promising renewable resource; yet, its operational and maintenance expenses pose significant challenges, particularly in offshore facilities. Proactive problem detection and predictive maintenance procedures are essential to enhance the reliability and cost-effectiveness of wind turbine systems. This research introduces a data-driven prognostic approach for forecasting the Remaining Useful Life (RUL) of wind turbine gearboxes, characterized by few failures but significant downtime and repair intricacy. The technology combines vibration and oil temperature data to develop robust health indicators through time-domain feature extraction and Principal Component Analysis (PCA). The Exponential Degradation Model is utilized for Remaining Useful Life estimation because of its appropriateness for systems demonstrating slow, monotonic deterioration. The assessment of performance by conventional measurements reveals significant predictive capacity, especially in the context of late-stage deterioration. Vibration data provides marginally superior predictive accuracy relative to oil temperature owing to its enhanced sensitivity to mechanical deterioration. The suggested methodology can accommodate real-world noise and sensor constraints. Overall, the proposed system provides a lightweight, interpretable, and effective solution for predictive maintenance in wind energy systems, with the potential to reduce unplanned downtime and improve operational reliability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data driven approach for estimation of remaining useful life of gearbox in wind energy conversion systems

  • 期刊论文

  • 发表日期：

    2025年6月13日

查看全文
  • R. Akash,
  • J. Preetha Roselyn,
  • U. Sowmmiya,
  • D. Devaraj

摘要

Wind energy represents a promising renewable resource; yet, its operational and maintenance expenses pose significant challenges, particularly in offshore facilities. Proactive problem detection and predictive maintenance procedures are essential to enhance the reliability and cost-effectiveness of wind turbine systems. This research introduces a data-driven prognostic approach for forecasting the Remaining Useful Life (RUL) of wind turbine gearboxes, characterized by few failures but significant downtime and repair intricacy. The technology combines vibration and oil temperature data to develop robust health indicators through time-domain feature extraction and Principal Component Analysis (PCA). The Exponential Degradation Model is utilized for Remaining Useful Life estimation because of its appropriateness for systems demonstrating slow, monotonic deterioration. The assessment of performance by conventional measurements reveals significant predictive capacity, especially in the context of late-stage deterioration. Vibration data provides marginally superior predictive accuracy relative to oil temperature owing to its enhanced sensitivity to mechanical deterioration. The suggested methodology can accommodate real-world noise and sensor constraints. Overall, the proposed system provides a lightweight, interpretable, and effective solution for predictive maintenance in wind energy systems, with the potential to reduce unplanned downtime and improve operational reliability.